MCML Junior Member Nikita Araslanov has been awarded a prestigious Emmy Noether Grant by the German Research Foundation (DFG).

The grant will support Nikita in establishing his independent research group at TUM, focusing on methods of Artificial Intelligence. The group is expected to launch by end of 2026 / early 2027, with PhD openings to be announced in due course.

Emmy Noether Grants are competitive funding awards that help early-career researchers start their own research groups and carry out independent projects.

Congratulations to Nikita!

Imagine applying for a job and being interviewed not by a human recruiter, but by an AI avatar on your screen. The conversation feels surprisingly natural. The interviewer smiles, asks questions, and responds to your answers in real time. But after the interview, the system rejects your application. Would it matter what the AI interviewer looked like?

The GRaM Competition 2026 has been launched as a benchmark challenge in conjunction with the Workshop on Geometry-grounded Representation Learning and Generative Modeling (GRaM).

Participants are invited to model airflow around 3D airfoil geometries inspired by a Formula 1 front wing. The task is to predict future airflow dynamics based on past velocity fields and the underlying geometry, using approaches such as neural operators, graph neural networks, or transformers.

The dataset, provided by BeyondMath, includes simulated airflow across a wide range of geometric configurations, enabling research on sequential point-cloud prediction.

Submissions will be evaluated on a held-out test set. Participants can contribute individually or in teams and submit their solutions via GitHub. All valid submissions will be included in the workshop proceedings, with the option for co-authorship.

The winner will receive the MCML Award, including 500 € in prize money.

The application deadline is April 22, 2026 (AoE).

Further details are available on the official competition website.

MCML PI Benjamin Busam researches how machines can perceive and understand their environment in three dimensions. At the intersection of computer vision, robotics, and artificial intelligence, he develops methods for 3D reconstruction from image and sensor data and creates digital twins to produce photorealistic and geometrically precise representations of our world. In this process, multimodal data is analyzed automatically.

Busam studied Mathematics and Physics at the Technical University of Munich (TUM), ParisTech in France, and the University of Melbourne. He earned his doctorate in computer science with a focus on 3D computer vision and led research groups in the industry as Head of Research at Framos and 3D Computer Vision Team Lead at Huawei Research in London. He has held the Professorship of Photogrammetry and Remote Sensing at TUM since September 1, 2025.

From April 13–17, the MCML will travel to the United Kingdom for a delegation visit aimed at further advancing its international collaboration in AI. This year’s visit focuses on leading academic institutions in London, Oxford, and Cambridge.

The trip brings together MCML researchers and partners for a week of exchange, networking, and joint scientific activities. Its overarching goal is to deepen existing partnerships, initiate new collaborations, and further position Munich as a leading European hub for AI and machine learning.

In this episode of #digdeep the digital expert Ralf Klüber talks about AI assistants and their meaning for executives and companies.

The podcast describes the rapid progress of AI agents, which now have direct access to personal data such as emails, files, and calendars and can perform tasks independently. New tools like Claude Code and OpenClaw make it possible to run such assistants locally on your own computer and link various information sources.

Technologies like the MCP protocol also promote communication between different AI systems, while platforms like Moldbook create new networks for AI agents.

In the interview, Ralf Klüber explains how he quickly built a personal AI assistant that combines several digital tools. Finally, the episode raises questions about the future of companies, the security of such systems, and the role of executives in dealing with AI.

Ralf Klüber is an experienced digitalization expert and former CTO at umlaut, who now lives as a full-time traveler and shares his experiences with technology and self-organization. He combines engineering, technology strategy, and entrepreneurship and, together with Frauke Kreuter and Christof Horn, brought DataFest to Germany.

As an active thought leader, he writes and speaks about AI, automation, and societal change, and documents his experiences on his blog, travel-fomo.com.

Just listen in (podcast in German).

The vacuum cleaner robot does the rounds, the heating regulates itself automatically and the lighting responds to an app. Every new device makes private households more convenient – and less secure. Under the aegis of ForDaySec, a Bavarian research association that is soon to be wound up, MCML Associate Johannes Kinder, chair of Programming Languages and AI at LMU, investigates how digitally secure our everyday life is and how connective devices can – not just in the short term – be guarded against undesirable access.

We are proud to share that the article “ChatGPT makes medicine easy to swallow: an exploratory case study on simplified radiology reports” has been awarded the Most Cited Article in European Radiology (Impact Factor 2024) by the European Society of Radiology.

In March 2026, the Munich Center for Machine Learning participated in a Bavarian economic delegation to the San Francisco Bay Area. Under the theme "Applied AI and the AI Ecosystem," Thomas Meier, Science Manager for Collaborations and PhD Coordinator at MCML, represented the center in a series of institutional visits and expert exchanges with leading organizations at the forefront of artificial intelligence research and application.

The MCML has reached an h-index of 100! For us, this represents many years of collaborative research and a strong partnership between LMU Munich and TUM. It reflects the contributions of our members, as well as the broader network of collaborators.

Thank you to everyone who has contributed along the way!

Machine learning models are good at giving answers, predicting patterns, and classifying objects but they are much worse at saying when they are unsure. But what if the model could admit when it was guessing and quantify the uncertainty?

In this episode of #digdeep MCML PI Sven Nyholm explains the improvements and convenience with AI solutions.

Who bears responsibility or credit when decisions or outcomes originate from AI systems? The rapid development of AI raises fundamental philosophical questions about what constitutes humanity when machines take over abilities such as thinking, writing, or decision-making.

Philosopher Sven Nyholm explains that classical ethical theories from thinkers like Aristotle, Immanuel Kant, and Jeremy Bentham are insufficient to evaluate modern AI phenomena such as autonomous systems or digital twins.

In this interview, he discusses risks and emerging questions surrounding responsibility, social interactions between AI agents, and the potential loss of human capabilities through increasing automation.

The interview also touches on topics such as AI regulation in Europe, Anthropic's unusual recruitment practices, and the concern that while AI might make life more convenient, it could also make it more susceptible to human laziness.

Just listen in (podcast in German).

Each year, the association honors researchers whose “contributions to the advancement of science or its applications are scientifically or socially outstanding” with this internationally renowned award.

MCML PI Christoph Kern and MCML Junior Member Jan Simson received the Best Paper Award at the GOR 2026 Conference in Cologne together with Fiona Draxler and Samuel Mehr.

The award-winning paper, “Preventing Harmful Data Practices by using Participatory Input to Navigate the Machine Learning Multiverse,” highlights the importance of making key decisions across the machine learning pipeline transparent and accessible to the public. The authors propose a participatory approach to navigating different design choices in machine learning and argue that central decisions should be democratized rather than purely optimized.

Congratulations from us!

MCML PI Enkelejda Kasneci and MCML Junior Member Ka Hei Carrie Lau have received an Honourable Mention Award at CHI 2026 for their paper “Skin-Deep Bias: How Avatar Appearances Shape Perceptions of AI Hiring”.

The award acknowledges the paper’s originality, rigor, and potential impact in human-computer interaction. The study investigates how avatar appearances affect perceptions in AI hiring, providing key insights into bias and fairness in AI systems.

Congratulations from us!

MCML PI Alena Buyx was recently featured in Süddeutsche Zeitung discussing the ethical boundaries of artificial intelligence.

Buyx explored questions such as whether AI should make life‑and‑death decisions. The discussion also included philosopher Julian Nida-Rümelin, who warned about the dangers of technolibertarianism in the USA.

Her insights contribute to the broader public debate on responsible AI use and ethical safeguards.

MCML PI Matthias Niessner was recently featured in an article on heise online about the rapidly growing European AI start-up ecosystem.

The article discusses how international investors continue to invest heavily in European AI companies despite concerns about a potential AI bubble. While generative AI still faces challenges such as hallucinations and limited world understanding, many start-ups are demonstrating promising applications and attracting significant venture capital.

MCML PI Reinhard Heckel was recently featured in Frankfurter Allgemeine Zeitung (FAZ) in the article “Wieso es in der Künstlichen Intelligenz jetzt erst recht auf hochwertige Daten ankommt.” The piece underscores why high‑quality data are crucial for training effective and fair AI systems, highlighting how better data improves model performance and helps mitigate bias and other pitfalls in AI.

From hyper-realistic video generation to protein design, Diffusion Models are the engine behind the current wave of Generative AI. But if you try to understand how diffusion models actually work, you quickly run into heavy math. Most explanations focus only on images. If you're working with text, biology, or graphs, you’re often left alone with scattered papers and technical jargon.

MCML PI Matthias Althoff was mentioned in VentureCapital Magazin in the context of the Munich robotics startup Sitegeist.

Founded in 2025 as a spin-off of the Technical University of Munich, the company originates from the robotics research environment associated with Althoff’s chair. The startup brings together expertise in AI, robotics, and infrastructure technology to develop autonomous robotic systems capable of repairing damaged concrete structures directly on construction sites.

The technology aims to improve efficiency in infrastructure maintenance and address labor shortages in the construction sector by enabling AI-driven robots to independently detect and remove deteriorated material.

MCML PI Sven Nyholm was recently featured in Big Think. In the article “The hidden cost of letting AI make your life easier,” Nyholm reflects on how increasing reliance on AI might affect our sense of achievement and meaning in life.

He argues that while AI can make many tasks easier, outsourcing effortful activities to machines may reduce opportunities for developing skills, exercising judgment, and experiencing genuine achievement. Nyholm highlights the importance of balancing technological assistance with meaningful human engagement, emphasizing that effort, creativity, and critical thinking remain central to a meaningful life.

In this episode of #digdeep Letitia Parcalabescu talks about how to explain AI terms in a coffee break.

AI systems' responses depend not only on developers but also on so-called system prompts, which determine how a model reacts. AI researcher Letitia Parcalabescu tries to make such technical details understandable because public understanding of AI often lags behind its widespread use.

She is particularly well-known for her YouTube channel "AI Coffee Break," where an animated coffee bean explains current AI research in short, easy-to-understand videos. In addition to her role as a science communicator, she works as a researcher at Aleph Alpha Research, focusing on interpretable reasoning models.

Her goal is to make complex AI topics more accessible and to show that science can also be communicated in a more engaging way.

Just listen in (podcast in German).

That the Universe is expanding has been known for almost a hundred years now, but how fast? The exact rate of that expansion remains hotly debated, even challenging the standard model of cosmology. A research team at LMU, TUM, and the Max Planck Institutes for Astrophysics (MPA) and for Extraterrestrial Physics (MPE) has now imaged and modelled an exceptionally rare supernova nicknamed SN Winny that could provide a new, independent way to measure how fast the Universe is expanding.

Yesterday’s MCML Stammtisch brought the community together — a perfect chance to exchange ideas and enjoy relaxed conversations beyond the labs.

Today’s AI systems are remarkably good at recognizing what stands out most. Yet understanding how details relate to each other remains surprisingly difficult. Imagine asking an AI assistant to describe a crowded street photo. It quickly mentions the car in the foreground — but misses the cyclist waiting at the light or the small sign that changes the meaning of the moment.

MCML Junior Research Group Leader Benjamin Lange was featured on the Bayerischer Rundfunk podcast IQ – Wissenschaft und Forschung in the episode “Kann KI lügen? – So manipulativ sind Chatbots.”

In the discussion, he addresses whether AI systems can intentionally deceive or merely produce false information due to technical limitations. He explains how so-called “hallucinations” in chatbots occur and clarifies common misconceptions about AI “lying.”

The episode also explores potential risks of manipulative AI systems and reflects on how society, research, and policymakers should respond to these challenges.

In this episode of #digdeep Jochen Kuhn and Florian Karsten explore how school will change when children grow up with AI and chatbots that can provide knowledge at any time and even take over tasks.

This raises the fundamental question of whether we are increasingly outsourcing our own thinking to machines. At the same time, the traditional concept of education is faltering because it is becoming increasingly difficult to distinguish between what comes from humans and what comes from AI.

In conversation with educational researchers it becomes clear that AI is now an integral part of both schools and children's bedrooms. At the same time, it becomes apparent that there are currently more open questions than clear answers regarding the use of AI in the education system.

Just listen in (podcast in German).

MCML Director Daniel Rueckert and MCML PI Fabian Theis have received funding through the $20 million AI for Science fund by Google.org. Their project, based at Technische Universität München and Helmholtz Munich, is among twelve selected worldwide.

The team will develop a multiscale foundation model that bridges individual cells and whole organs, enabling clinicians to digitally simulate disease progression and test treatments.

Long videos are everywhere, with footage of movies, YouTube videos, body-cam recordings, and AR/VR often running for tens of minutes or even hours. Now imagine asking a simple question like “Where are the scissors in this video?”

MCML PI Benjamin Busam is leading the scientific design of the Bavarian Earth observation satellite network “CuBy”, which is scheduled for launch in June 2026.

The system consists of five small satellites that will jointly acquire high-resolution, multi-spectral data covering the entire state of Bavaria. The project is implemented entirely in Bavaria, from data acquisition in space to ground-based processing and public data provision.

The satellite data will support applications in agriculture, environmental monitoring, flood prediction, and civil protection. Combined with AI-based data analysis and management systems, “CuBy” strengthens Bavaria’s position as a key location for Earth observation and data-driven decision-making.

There was a time when the Andromeda Nebula was the end of the Universe. At least, a number of astronomers in the early 20th century still thought that fuzzy objects in the night sky – such as the Andromeda Nebula – were part of our Milky Way, and that our Galaxy was all there was to the Universe. Other researchers saw the nebulae as distant stellar islands, believed that the space in-between was barren and empty – and held the Universe to be considerably bigger. Observation data were unable to clear the matter up: Back then, it was not possible to measure the distance to the stars. Today, we know that the Andromeda Nebula is a separate galaxy that is 2.5 million light years away.

In this episode of #digdeep the team speaks with Anna Kopp, Director of Microsoft Digital Germany, about the meaningful integration of AI, agents, and human work.

She describes how work at Microsoft has changed dramatically over the past 20 years, especially since the pandemic, due to more meetings, emails, and increased complexity.

The discussion explores why AI has become indispensable for many knowledge workers, but how its effectiveness depends heavily on governance, data quality, and corporate culture. A key focus is on agentic AI and how companies can manage thousands of AI agents in a structured way, agents that function more like new colleagues than mere apps.

Furthermore, it is emphasized that successful transformation requires a digital mindset, a learning culture, leadership responsibility, and time for reflection.

Just listen in (podcast in German).

How can machines learn the hidden rules that govern how systems change—how objects move, how weather patterns unfold, or how biological signals evolve—without ever being told what those rules are?

The London-based AI company Synthesia, co-founded by MCML PI Matthias Nießner, has raised an additional $200 million, bringing its valuation to $4 billion. Synthesia is one of Europe’s most valuable AI startups and a pioneer in AI-generated video avatars for corporate training and communication.

With the new funding, Synthesia aims to expand beyond static training videos toward interactive AI avatars that can simulate conversations, test employees’ knowledge, and provide personalized feedback. These virtual trainers are designed to support onboarding, sales training, and internal education, enabled by recent advances in generative AI and large language models.

MCML PI Fabian Theis was featured in an interview with FOCUS online at DLD Munich 2026, discussing how artificial intelligence can transform medicine.

Fabian highlighted how AI can help make diagnosis and treatment more accessible, efficient, and personalized, for example by supporting the interpretation of complex medical data such as blood tests. He also shared his vision of a “Virtual Human” as a digital model spanning molecules, cells, organs, and the whole body to enable better predictions and personalized care.

Rather than replacing doctors, AI should support them by reducing administrative workload and giving them more time for patients. Used responsibly, Fabian emphasized, AI can empower both medical professionals and patients alike.

In January 2026, two research groups of the MCML and Tübingen AI Center met for a four-day retreat in Gaschurn/ Austria. About 60 researchers joined for talks and poster sessions. The joint retreat highlighted the latest advances in Computer Vision, driven by foundational models, generative AI, and 4D reconstruction and deformation methods.

A central topic was 3D and 4D spatial reasoning, from the macro-scale of aerial UAV localization through orthographic geodata, to the visual-interial tracking of hands and humans. Together, these efforts reflect a push toward robust spatial intelligence that works in broad real-world settings.

Our Director Daniel Rückert spoke at the DLD Conference 2026 as part of the BIOSPHERE Health Track.

In his talk, Daniel explored how AI is moving from science fiction into real clinical practice and what this transformation means for the future of medicine. The session is now available in the DLD Conference video series.

We are pleased to share that our PI Fabian Theis contributed to DLD Conference 2026 as part of the BIOSPHERE Health Track.

Fabian gave a talk titled “From Models to Medicines: AI-guided Experimental Biology”, focusing on how AI supports biological research and drug discovery. In addition, he joined the panel discussion “From the Cell to Clinical Application”, addressing the translation of AI-driven insights from basic research into clinical practice.

Both sessions are now available as part of the DLD Conference video series.

Our PI Björn Ommer was a featured speaker at DLD Munich 2026 in the session “It’s Gonna Be Wild: When AI Moves Faster Than Society”, together with LMU VP Armin Nassehi.

The discussion reflects on the growing gap between rapid AI development and society’s ability to adapt. The session is now available as part of the DLD Conference video series.

On January 20th, we visited DeepL at their Munich office for an edition of our “MCML Pitchtalks with Industry.”

DeepL presented their current state-of-the-art work in NLP, and our junior members—Siyao Peng, Daryna Demientieva, Felicia Körner, and Ivica Obadic—showcased their cutting-edge research on value alignment, translation challenges, problematic speech detection, and vision–language models.

MCML looks forward to building further collaborations with DeepL.

Machine learning models can be powerful, but understanding why they behave the way they do is often much harder. Early global interpretability tools were designed to show how each feature affects the model’s final outcome. But these tools typically work by averaging over the entire dataset, however averages can often hide important structure.

Between Freudenberg – “happiness mountain” – and Rosenberg – “roses mountain”, I had the pleasure to visit the Institute of Computer Science (ICS-HSG) at the University of St. Gallen (HSG) in Switzerland for two weeks, as part of the MCML AI X-Change Program, while being a PhD candidate at LMU Munich and the Munich Center for Machine Learning (MCML). I am deeply grateful to Prof. Dr. Barbara Weber and her Software Systems Programming and Development Chair for hosting me, as well as to the wonderful team at ICS-HSG — including Marco Franceschetti, Lisa Zimmermann, Ronny Seiger, Hagen Völzer, Amine Abbad-Andaloussi, Aaron Kurz, and Thierry Sorg — for the inspiring discussions, warm welcome, and open exchange of ideas.

MCML PI Matthias Nießner was featured in WirtschaftsWoche for his startup Spaitial AI, which develops AI capable of creating realistic 3D models of real-world rooms and interiors within seconds. The article highlights Nießner’s work as a pioneering effort in spatial AI, showing how this technology could transform architecture, virtual design, and digital content creation. Spaitial AI is gaining attention as a potential next success story in Germany’s AI landscape.

At the first GirlsLAB at TUM, female students from Bavarian partner schools explored STEM subjects through hands-on teamwork and creative problem-solving. While the students built chain-reaction machines on campus, accompanying teachers took part in a workshop co-designed by the Munich Center for Machine Learning and KI macht Schule (“AI goes to school”) on using AI tools in the classroom. The program offered insights into university life and encouraged girls to explore future paths in science and technology.

MCML PI Frauke Kreuter was featured on ARD alpha, discussing how AI is shaping everyday life and the workplace. Alongside moderator Birgit Kappel and Sabine Pfeiffer, head of the Chair of Sociology at FAU Erlangen-Nuremberg and director of the Bavarian Research Institute for Digital Transformation (bidt), she explored current AI capabilities, future prospects, and the importance of responsible AI use, highlighting insights from her international research in social data science.

In this episode of #digdeep, Christof Horn reports live from the CES, the world's largest consumer electronics trade show in Las Vegas and now also one of the leading trade shows for the global automotive industry.

While there were hardly any classic new car models on display at CES, profound technological changes are taking place behind the scenes.

The focus is shifting from visible hardware gimmicks to fundamental AI structures such as training data, simulations, reasoning models, and scalable infrastructures. Automated driving remains the central AI application in cars, particularly in handling rare, safety-critical "long-tail" situations. "AI for AI," for example through simulation and training, plays a crucial role here, as demonstrated by NVIDIA, among others.

The overarching theme of CES was "Physical AI," where AI models conquer the physical world and autonomous vehicles represent only one part of a future environment shared by intelligent systems.

Just listen in (podcast in German).

Vision-language models have shown that images and text can live in a shared space: a picture of a "cat" often lands close to the word "cat" in the embedding space. But such models learn this correspondence because they have seen millions of image–caption pairs. What if we had no paired data at all? Could we still figure out which images match which words?

MCML PI Alexander Fraser was featured as a guest on an episode of the Breaking Data podcast, where he discussed the importance of integrating low-resource and minority languages into AI models. In conversation with host Nico Litzel, Fraser talked about machine translation in niche domains, intellectual property challenges, and the societal benefits of chatbots for low-resource languages.

MCML PIs Xiaoxiang Zhu and Felix Dietrich were featured in Süddeutsche Zeitung’s coverage of TU Munich’s major scientific breakthroughs in 2025. Their work contributes to current advances in data-driven methods and AI research and was highlighted as part of TU Munich’s research activities in 2025.

The certificate course "Responsible AI" is jointly organized by the German National AI Centers. It addresses AI researchers on all levels from the centers, the participating universities, as well as experienced AI enthusiasts and users.

Image generation models today can create almost anything, like a futuristic city glowing at sunset, a classical painting of your cat, or a realistic spaceship made of glass. But when you ask them to go bigger and sharper, the magic slows down. The process takes longer, eats up more memory, and trying out new versions can feel endless. For quite long, high-resolution AI images have come with an annoying trade-off: do you want speed, or do you want perfection?

Konstantin Riedl, former MCML member, has received the Doctoral Award of the Friends of TUM e.V., honoring one of the most outstanding doctoral dissertations at the Technical University of Munich.

He is currently a Postdoctoral Research Associate at the University of Oxford and a Research Fellow in Artificial Intelligence & Machine Learning at Reuben College. The award recognizes his scientific excellence and his contributions at the intersection of mathematics and modern technologies.

Congratulations to Konstantin Riedl on this well-deserved achievement.

In this episode of #digdeep, the sociologist Armin Nassehi declares, how social transformation succeeds.

He discusses, why social changes are often not implemented despite clear goals and great urgency and explains this with a collective transformation fatigue, which is not an individual failure, but is rooted in the structures of modern societies.

A central point is the distinction between organizations that function through clear control logics and societies that cannot be managed according to the same principles. According to Nassehi, real transformation comes less through large political gestures or moral appeals than through laborious, small-scale negotiation processes and cooperation between different actors.

Finally, AI is classified as a profound media disruption that questions classic ideas about expertise and requires new forms of competent handling of knowledge.

Just listen in (podcast in German).

Thank you for a wonderful year of collaboration and partnership.

Wishing you a joyful festive season, and we look forward to seeing you again in 2026!

Teachers need support to keep up with rapid technological change and to adapt teaching and learning to the demands of an increasingly digital world. While new technologies are widely available, they are often used only as additional tools within familiar classroom structures. Simply introducing AI into education does not, by itself, lead to new ways of learning. The KITrans project therefore focused on a central question: how can teachers be effectively supported in using AI in teaching and learning?

To address this, the MCML together with its project partners— the Chair of Primary School Education and Didactics at LMU Munich and Junge Tüftler gGmbH—worked closely with teachers from different types of schools. Together, they developed and tested AI-related teaching and learning approaches, which are now available on Junge Tüftler gGmbH’s “TüftelLab digital” platform. A key focus was on practical materials that teachers can easily use in their everyday classroom practice.

In 2025, MCML expanded its outreach activities with the aim of supporting students and teachers in developing a clearer understanding of artificial intelligence and its real-world applications. Through workshops for school classes, participants were introduced to fundamental questions such as how computers recognize patterns, how machine learning differs from traditional programming, how models interpret images, and how AI is used in medical diagnostics.

Throughout the year, MCML Junior Members conducted workshops for school classes at different educational levels in Bavaria and beyond, bringing current AI research and applications directly into the classroom.

MCML has launched the Thinkathon, a new initiative designed to connect AI research with the practical needs of the Bavarian education system. In a hackathon-style setting, MCML Junior Members collaborated with mentors and stakeholders from the Bavarian Ministry of Education and other partner organizations to develop hands-on, AI-based teaching kits for real-world classroom use.

Using AI and LLMs at work feels almost unavoidable today: they make things easier, but they can also go wrong in important ways. One of the trickiest problems? Gender bias. For example, if you ask about someone’s skills from a photo, it may confidently label them a "born leader" or "working well under pressure" with no real evidence.

Our PI Fabian Theis was recently featured in an interview exploring how artificial intelligence is reshaping modern healthcare. As a leading researcher in AI-driven biomedical data analysis, he explains why precision medicine is still widely underestimated, despite major scientific advances. AI now enables more individualized predictions and treatment strategies, with developments increasingly making their way from research into clinical practice. Theis also reflects on current digital health trends, emphasizing the need to distinguish short-term hype from technologies with lasting clinical impact.

Our PI Gitta Kutyniok was featured in VDI Nachrichten discussing ethical and safety challenges for autonomous AI agents. She emphasizes the need for measurable criteria to guide AI decision-making, ensuring responsible and transparent behavior. Kutyniok’s insights highlight how rigorous mathematical frameworks can support trustworthy AI in real-world applications.

Our PI Hinrich Schütze commented on the Swiss start-up Giotto.ai’s breakthrough in language models, featured in WirtschaftsWoche. Giotto.ai challenges the prevailing trend of ever-larger AI models by developing a highly efficient 200-million-parameter system that leverages external memory and context-aware decoding to reduce errors and hallucinations. Schütze highlights that separating memory from the model addresses fundamental inefficiencies in current AI systems, making them more robust and capable.

Our PI Xiaoxiang Zhu and her team have created a Global Building Atlas, mapping 2.75 billion buildings in 3D. Using satellite data and AI, the atlas captures building footprints, geometry, and height, providing insights into urbanization, housing, and infrastructure. The project introduces building volume per capita as a new indicator for living space and social inequalities, and supports planning for energy, environment, and disaster management.

In the new episode of #digdeep, Sabine Pfeiffer from FAU Erlangen-Nuremberg talks about AI as a new colleague.

An MIT study shows that artificial intelligence is widespread in everyday life, but lags behind expectations in business. Optimistic narratives about AI are of little help when it comes to actually using the technology productively.

Industrial applications still require considerable expertise to develop marketable products and services. For AI to increase productivity and automate routine tasks, humans and machines must work closely together, as human judgment remains essential.

The success of AI implementations in companies depends crucially on the early involvement of employees, as Professor Sabine Pfeifer also emphasizes in her research.

Just listen in (podcast in German).

Ever wondered how a 3D shape can smoothly change — like a robot arm bending or a dog rising from sitting to standing — without complex simulations or hand-crafted data? Researchers from MCML and the University of Bonn tackled this challenge in their ICLR 2025 paper, "Implicit Neural Surface Deformation with Explicit Velocity Fields".

A group of our Junior Members visited Munich Re for the Latest Edition of Pitchtalks with Industry. We had the pleasure of visiting Munich Re for another edition of MCML’s Pitchtalks with Industry.

MCML JRG Leader Tom Sterkenburg has been awarded the Karl-Heinz Hoffmann Prize of the Bavarian Academy of Sciences and Humanities (BAdW). The prize was presented at the Academy’s Ceremonial Annual Meeting on 6 December 2025 in Munich.

With its annual science prizes, the BAdW honors outstanding early-career researchers across disciplines. The Karl-Heinz Hoffmann Prize is awarded alternately in the humanities and the natural sciences. Tom Sterkenburg’s research is situated at the intersection of philosophy, statistics, and computer science. His work combines mathematical modeling, algorithmic simulation, and philosophical analysis to address the classical problem of induction in the context of machine learning.

Congrats to Tom on this achievement!

Our PI Björn Ommer was a guest in the ZDF science series Terra X - moderated by Harald Lesch - where he explained how image-generative AI works and how systems like Stable Diffusion enable machines to “see” and create images.
In the episode Björn Ommer discusses the technological foundations of generative models and their impact on art, creativity, and society.

Our PI Björn Schuller was a guest in the ZDF science series Terra X - moderated by Harald Lesch, where he discussed the fundamentals of Affective Computing and the role of emotion recognition in artificial intelligence.
The episode explores how AI systems can detect emotions from speech, facial expressions, and text, as well as the ethical challenges surrounding empathic AI.

The MCML held its annual General Assembly last week, bringing its members together for networking and discussion. The program opened with updates and highlights from the MCML directors, followed by an overview of the center’s workshop and training activities.

With the GlobalBuildingAtlas, the research team led by MCML PI Xiaoxiang Zhu has created the first high-resolution 3D map of all buildings worldwide. The open data provides a crucial basis for climate research and the implementation of the UN Sustainable Development Goals. They enable more precise models for urbanization, infrastructure and disaster management – and help to make cities around the world more inclusive and resilient.

Large language models like ChatGPT or Gemini are now everywhere, from summarizing text to writing code or answering simple questions. But there’s one thing they still struggle with: admitting uncertainty. Ask a fine-tuned LLM a tricky question, and it might sound quite confident, even when it’s completely wrong. This "overconfidence" can be risky in areas such as healthcare, finance, and science, where trust and uncertainty matter. As these tools become daily information sources, many people treat them as unquestionable authorities, skipping fact-checking or deeper research. The danger isn’t just that AI might be wrong now, but that its confident mistakes can spread, reinforce, and grow, teaching future models what "truth" sounds like.

For more than 30 years, the Bavarian State Center for Political Education has been organizing the “Lernort Staatsregierung” program in the Bavarian State Ministries. School classes from all types of secondary schools across Bavaria are invited to take part.

The program offers young people a personal and vivid insight into the tasks, structures, and working practices of the Bavarian state government. As part of their visit to the Bavarian State Ministry for Science and the Arts (StMWK), the MCML contributed a talk on AI. MCML Junior Members introduced students to current AI research, real-world applications, and key societal and political questions surrounding the use of AI. Around 30 students from the Q12 of Dante Gymnasium in Munich attended. MCML’s contribution highlighted the growing role of AI in science, public administration, and society, and emphasized the importance of informed and critical engagement with these technologies in a democratic context.

From May to July, I spent three exciting months as a visiting researcher at the Computer Science Department of Princeton University, hosted by Prof. Manoel Horta Ribeiro. The visit grew out of a keynote Manoel gave at LMU. After his talk, we discussed potential joint projects at the intersection of causal inference, machine learning, and social computing, which eventually developed into this research stay. I am very grateful to the MCML AI X-Change program, to Manoel for the kind invitation and continuous support, and to everyone at Princeton, LMU, and MCML who made this stay both productive and enjoyable.

MCML PIs Daniel Grün and Lukas Heinrich are part of the newly funded Germany-wide project “SciFM,” supported by the BMFTR, which aims to develop foundation models for astro- and particle physics.

Modern particle colliders, such as the Large Hadron Collider at CERN, and telescopes like the Square Kilometer Array generate exabytes of data, creating an ideal testbed for AI. SciFM seeks to create a new generation of advanced AI tools—so-called foundation models—that act as “AI generalists,” trained on diverse datasets to solve multiple tasks, outperforming models trained on a single experiment.

The project unites leading researchers in AI for fundamental physics across four locations: Munich, Heidelberg, Hamburg, and Aachen, and has received funding of over 3 million euros. SciFM is currently hiring postdoctoral researchers and PhD students.

Four members of the MCML have been selected for the GAIE Spotlight Series: Leading 30 German AI Experts & Academics 2025. The list recognizes outstanding researchers shaping the future of artificial intelligence in Germany and beyond.

Large vision-language models (VLMs) like CLIP (Contrastive Language-Image Pre-training) have changed how AI works with mixed inputs of images and text, by learning to connect pictures and words. Given an image with a caption like "a dog playing with a ball", CLIP learns to link visual patterns (the dog, the ball, the grass) with the words in the caption. This lets it recognize and describe new images it has never seen before, a skill known as zero-shot learning.

The InterACT Workshop 2025 took place at LMU Munich in September. Participants engaged in an intensive scientific exchange on methodological challenges in Explainable AI (xAI). The workshop brought together 24 researchers from eight institutions across Europe: LMU Munich, University of Bremen, University of Warsaw, University of Zurich, Norwegian Computing Center, University of Tübingen, Universidad Rey Juan Carlos Madrid, and Universidad Internacional Menéndez Pelayo Madrid. The workshop focused on Interpretable and Explainable Machine Learning (IML/xAI) and was organized by Fiona K. Ewald (LMU & MCML) with support from MCML Director Bernd Bischl.

The AI in Medicine Workshop 2025 brought together more than 120 experts from medicine and computer science at the Baltic, Charité’s innovation hub for digital health.

MCML Director Daniel Rückert is listed among the world’s most frequently cited researchers in the Cross-Field category for his work on artificial intelligence in healthcare and medicine.

In total, 17 TUM scientists were recognized in the 2025 Highly Cited Researchers rankings by Clarivate, highlighting their significant contributions to their respective fields.

During my research stay at Stanford University from July to September 2025, I had the pleasure of being part of the research group led by Assistant Professor Serena Yeung in the Department of Biomedical Data Science. My two-month stay in California gave me the opportunity to investigate how public scientific articles can be leveraged to build biomedical vision-language foundation models, culminating in a paper accepted to ML4H. Together with the group, I also laid the groundwork for several projects.

In the new episode of #digdeep, Moritz Mayer talks about the shape product development of Synera.

Developing physical products requires specialized knowledge and the use of complex tools. Moritz Maier and his startup Synera want to simplify this process by creating a platform that enables product development such as “low code” programming.

On this platform, products can be developed using a visual language of simple elements and functions, with AI agents taking over routine tasks. This approach creates a consistent data environment that meaningfully supports the use of artificial intelligence.

Despite difficult initial phases, the team is now striving for international success, especially in the USA.

Just listen in (podcast in German).

How can components be designed for an optimal balance of minimal weight and maximum robustness? This is a challenge faced by many industries, from medical device manufacturing to the automotive and aeronautics sectors. MCML PI Rüdiger Westermann and his team have developed a reference system that permits direct comparisons and evaluations of many different lightweight construction methods.

State-of-the-art diffusion models like DiT and Stable Diffusion have made AI image generation incredibly powerful. But they still struggle with one big issue: scaling to large images or videos quickly and efficiently without exhausting your GPU memory. What if we could process images faster, use less memory, and still retain visual quality—without sacrificing the model’s expressiveness?

The paper from MCML Junior Members Vincent Tao Hu, Olga Grebenkova, Pingchuan Ma, Johannes Schusterbauer, and MCML PI Björn Ommer, and collaborators Stefan Andreas Baumann, and Ming Gui, introduces ZigMa, a plug-and-play diffusion backbone that combines the strengths of DiT-style architectures and Mamba’s efficient long-range sequence modeling.

MCML PI Anne‑Laure Boulesteix has been named a Highly Cited Researcher 2025 by Clarivate for her influential work in the Cross-Field category.

The recognition highlights publications from 2014–2024 that rank among the top 1% most cited worldwide, reflecting Anne‑Laure Boulesteix’s global impact and the international visibility of her research. She is one of 17 LMU scholars recognized in this year’s Highly Cited Researchers list.

MCML PI Björn Ommer was recently featured in Frankfurter Rundschau, commenting on the impact of Google’s new AI-powered search mode. The article discusses how AI-generated summaries in search results may reduce traffic to publishers and reshape how users access information online.

Ommer emphasized that while this development offers greater convenience, it also raises important questions about transparency, fairness, and the visibility of journalistic content in an AI-driven information landscape.

MCML PI Fabian Theis is listed among the world’s most frequently cited researchers in Biology and Biochemistry for his work on mathematical modeling of biological systems.

In total, 14 TUM scientists were recognized in the 2025 Highly Cited Researchers rankings by Clarivate, highlighting their significant contributions to their respective fields.

Picture a typical day in a warehouse: one worker lifts, bends, and carries out the same task over and over again. While the routine may seem simple, the physical toll steadily builds—affecting joints and muscles. To combat the long-term health risks associated with such repetitive movements, businesses are increasingly turning to exoskeletons and exosuits—innovative wearable technologies designed to support the body and ease the burden of manual labor.

As part of the All Hands Meeting of the Network of German Centers of Excellence for AI Research in November 2025, the centers' working group on diversity, equity, and inclusion topics organized the workshop "Stronger Together. Building a Cross-Institutional Gender Equality and Diversity Movement in German AI Centers."

This year’s All Hands Meeting in Saarbrücken gathered researchers from all German AI centers (BIFOLD, ScaDS.ai, Lamarr Institute, Tübingen AI Center, DFKI and MCML) to exchange ideas and discuss current AI research.

On 4 November 2025, Stadtwerke München (SWM) welcomed representatives from the Munich Center for Machine Learning (MCML) for a focused session of Industry Pitchtalks. The event highlighted the growing synergy between academic research and real-world applications of artificial intelligence in energy, utilities, and process analytics.

On 29 October 2025, the workshop “AI, Ethics and Society” brought together researchers and practitioners from the Junges Kolleg | BAdW, Munich Center for Machine Learning (MCML) and bidt – Graduate Center für Postdocs to explore how artificial-intelligence algorithms shape individual and collective memory.

LMU and Harvard researchers are developing smarter and safer wearable technologies that adapt to the people using them. Their latest method not only optimizes how an exosuit supports workers during lifting, but also explains why these decisions are made—bringing transparency and human expertise into the process.

Tuning exosuits is a delicate task: engineers must find just the right balance of assistance for each person, often through trial and error. This is where Bayesian optimization (BO) helps—an AI approach that efficiently searches for the best settings. However, BO typically acts as a black box. To address this, MCML researchers Julia Herbinger, Yusuf Sale, and Giuseppe Casalicchio, together with MCML Director Bernd Bischl and PI Eyke Hüllermeier, contributed to ShapleyBO—a new framework that makes BO’s reasoning explainable and interactive.

The work, carried out in collaboration with the Harvard Biodesign Lab, shows how combining human insight and AI can lead to faster, safer, and more personalized exosuit technology.

The team of MCML PI Alexander Fraser together with researchers at TU Darmstadt have studied how text-to-image generators deal with gender stereotypes in various languages. The results show that the models not only reflect gender biases, but also amplify them. The direction and strength of the distortion depends on the language in question.

The paper “CSG-Fusion: Consistent Sparse-View Gaussian Splatting via Matching-based Fusion” by MCML Director Daniel Cremers and Junior Member Weirong Chen, together with Wenbo Ji, and Yan Xia, received the Best Paper Award at the ICCV 2025 Workshop on End-to-End 3D Learning.

CSG-Fusion reduces redundant Gaussian representations in pixel-aligned pointmaps for compact, consistent 3D reconstructions using a matching prior, mask-based feature aggregation, and context-view supervision.

Congratulations to the team on this outstanding achievement!

MCML PIs Barbara Plank and Hinrich Schütze were featured in a recent ARD Capriccio segment on how AI struggles to understand dialects. The broadcast highlighted research from the MaiNLP lab, with insights from MCML Junior Member Verena Blaschke, showcasing the challenges of applying AI language models to regional variations.

The paper "CoProU-VO: Combining Projected Uncertainty for End-to-End Unsupervised Monocular Visual Odometry" by MCML Director Daniel Cremers and Junior Members Weirong Chen and Johannes Meier, together with Jingchao Xie, Oussema Dhaouadi, and Jacques Kaiser received the Best Paper Award at GCPR 2025.

The Munich Career Fair for AI & Data Science took place on October 23, 2025, at TranslaTUM in Munich. Organized by MDSI, relAI, MCML, and the AI Hub@LMU, the event brought together students, researchers, and industry representatives to discuss career opportunities in AI, machine learning, and data science. Companies such as Celonis, DENSO, GE Healthcare, SAP, and ZEISS presented their work and career paths, while participants had the chance to network, exchange ideas, and explore potential collaborations in the field.

MCML Junior Member Unai Fischer-Abaigar, was featured in a recent episode of Executive Code. He discussed his paper "The Value of Prediction in Identifying the Worst-Off", exploring how governments use AI to allocate limited resources and when improving predictive models is more effective than simply expanding access to public programs. Using real data from Germany’s employment offices, his research challenges the assumption that better prediction always leads to better outcomes in public decision-making.

In the new episode of #digdeep, Anne Greul shows startup-pivots form herCareer.

Anne took the courageous step into entrepreneurship and learned early on that the path is rarely straightforward. Instead of pursuing a classic career in consulting or the automotive industry, she founded a startup in the Web 3.0 and blockchain sector with partners – but the idea didn't pan out.

After this setback, she started new with Leegle.AI and now uses artificial intelligence to help companies comply with complex regulations. Her story shows how important perseverance and the courage to start over are.

That's precisely why Anne was a guest on #digdeep – the podcast from the Brand Eins family, recorded at the herCareer trade fair in Munich, with the support of Clockodo.

Just listen in (podcast in German).

The MCML Stammtisch was a great chance for our members to connect across disciplines over drinks and relaxed conversations. As always, it was a casual setting to share ideas, discuss projects, or simply enjoy time together beyond the lab.

As of October 1, our PI Björn Ommer, has been appointed LMU’s Chief AI Officer.

In his new role, he will help strengthen LMU’s strategic profile in Artificial Intelligence and further expand research and transfer activities in this key area – in close collaboration with TUM and non-university partners.

The Data Science Certificate Program at LMU Munich, offered in collaboration with MCML, is a graded, part-time academic training program. It is designed for working professionals, researchers, and executives who want to improve their skills in the field of data analytics, statistics, and machine learning.

The program is offered twice a year. Apply now and join the training between March and June 2026!

Deep learning models are emerging more and more in everyday life, going as far as assisting clinicians in their diagnosis. However, their black box nature prevents understanding errors and decision-making, which arguably are as important as high accuracy in decision-critical tasks. Previous research typically focused on either designing models to intuitively reason by example or on providing theoretically grounded pixel-level and rather unintuitive explanations.

Successful human-AI collaboration in medicine requires trust and clarity. To replace confusing AI tools that increase clinicians’ cognitive load, MCML Junior members Tom Nuno Wolf, Emre Kavak, Fabian Bongratz, and MCML PI Christian Wachinger created SIC for their collaborators at TUM Klinikum Rechts der Isar. SIC is a fully transparent classifier built to make AI-assisted image classification both intuitive and provably reliable.

On October 7th, we hosted another session in our "Pitchtalks with Industry" series — this time in collaboration with Munich NLP. The event brought together more than 50 members of Munich’s NLP community.

On 25 September 2025, LMU Munich hosted the panel discussion “Ethics of AI Agents” at the Carl Friedrich von Siemens Stiftung, in cooperation with MCML, the appliedAI Institute for Europe, the TUM Think Tank, the Bavarian State Ministry for Digital Affairs, the Liquid Legal Institute, and the Technical University of Munich.

In the current episode of #digdeep, Cihan Sügür, associated partner of MHP, talks about digitalization in Germany.

Since 2024, Capital magazine has ranked Cihan Sügür among its "Top 40 under 40" individuals driving Germany forward with innovative strength. He is a prime example of people who break new ground with courage and determination.

Cihan is driving digitalization forward and gained experience at Porsche, Olympus, IBM, and Deutsche Bahn before becoming an Associated Partner at MHP.

As a winner of the CIO Young Talent Awards, a CIO Foundation scholarship holder, and an MBA student at WHU, he combines professional excellence with social commitment, including as a member of Atlantik-Brücke.

His motto: To shape Germany's future – with conviction, drive, and entrepreneurial spirit.

Just listen in (podcast in German).

Unai Fischer Abaigar is a researcher at MCML whose work focuses on improving decision-making in public institutions by developing AI systems that are both fair and effective in practice.

On September 24th and 25th, the first MCML-LAMARR-Workshop took place at the University of Bonn. Main focus was NLP, and its related areas. PIs and junior members from both German AI Centers came together to present each others research, to be part of networking sessions and to work in discussion groups, and to foster engagement in collaboration between our centers. PIs Lucie Flek and Mehdi Ali from LAMARR, and JRG Leader Michael Hedderich, alongside many PhD students, gave talks on their research. We all agreed on the fact that there is a lot of work to do in NLP with respect to LLM finetuning, alignment, and bias and ethics research.

We are happy to announce that the paper “Overtuning in Hyperparameter Optimization” by MCML Thomas Bayes Fellow Matthias Feurer, MCML Junior Member Lennart Schneider, and MCML Director Bernd Bischl has won the Best Paper Award at the 4th AutoML Conference.

The paper defines and studies overtuning, where hyperparameter optimization can improve validation scores but hurt generalization, and presents a large-scale analysis of factors influencing this effect. These insights will help the AutoML community design more robust and reliable optimization strategies

Congratulations to the team on this outstanding achievement!

Since September, MCML PI Alena Buyx has been co-hosting NANO Talk on 3sat, alternating with journalist Stephanie Rohde and following in the footsteps of renowned science communicator Gert Scobel.

The show brings science to life, discussing current and important topics with clarity, calm, and an interdisciplinary perspective.

Episodes covering AI and nutrition are already available in the media library and on YouTube, with many more exciting discussions to come.

How can machine learning fight climate change? Kerstin Forster, researcher at LMU and MCML, explores how AI can help reduce greenhouse gas emissions, improve renewable energy systems, and enhance early warning for extreme weather.

MCML PI Björn Ommer was featured in WELT discussing the challenges of AI neutrality and the risks of over- or under-correcting bias in large models. He highlighted that no system can be fully neutral since training data always reflects real-world imbalances, and emphasized the importance of maintaining human oversight in critical areas such as law and ethics.

Matthias Feurer is a Thomas Bayes Fellow and interim professor, funded by the MCML and a member of the Chair of Statistical Learning and Data Science at LMU. He aims to simplify the usage of machine learning by researching methods and developing tools that allow the usage of machine learning by domain scientists and also make machine learning more efficient for expert users.

MCML Principal Investigator Björn Schuller is mentioned in a recent Macwelt article discussing how the Apple Watch contributes to health monitoring and the future of healthcare.

He highlights the potential of smartwatches to detect subtle behavioral and physiological changes, to use audio and sensor data for early detection of conditions, and integrate multiple data streams for a more complete view of a user’s health. He also points out the importance of balancing data collection with user well-being.

Imagine re-running feature importance plots and getting slightly different “top features.” Annoying, right? That uncertainty often comes from a quiet assumption: model explanation algorithms typically sample points from data at random.

A new ICLR 2025 Spotlight paper by MCML Junior Member Giuseppe Casalicchio, MCML Director Bernd Bischl, first author Hubert Baniecki and co-author Przemyslaw Biecek flips this script with a simple idea: compress your data distribution first, then explain the model. The authors call it Compress Then Explain (CTE) and it consistently makes explanations more accurate, more stable, and much faster.

MCML PI Björn Ommer joined the ZDF programme NANO Talk, hosted by MCML PI Alena Buyx, to explore the question: What is the essence of artificial intelligence?

In the discussion, Ommer reflected on the technological foundations of AI, its societal and ethical implications, and how these developments may shape the future.

When machines don’t just provide answers but can also plan and even take action on their own, what does that mean for the people using them? Our Junior Research Group Ethics of AI, led by Benjamin Lange, studies the opportunities and threats of new AI agents.

Patients have more confidence in autonomous robotic ultrasound systems when an avatar guides them through the process. This was discovered by our PI Nassir Navab, Chair for Computer Aided Medical Procedures & Augmented Reality at TUM. The virtual agent explains what it is doing, answers questions and can speak any language. Such systems are intended especially for use in regions where there is a shortage of doctors.

How can AI predict medical conditions and personalize treatments? Simon Schallmoser, researcher at LMU and MCML, uses machine learning to forecast health risks and optimize care for patients based on their individual profiles.

MCML featured a workshop series specifically designed for primary school students. The Gemeindebücherei Eching, in collaboration with the MCML, hosted “Computers with Brains”: An AI Workshop Series for Primary School Students from the Munich area.

The workshops, led by Felicia Körner, a Junior Member from MCML, introduced the students to AI in a playful and accessible way. They experimented with AI-generated images using “KinderGPT” and practiced recognizing real vs. AI-generated content. The sessions also emphasized safe and responsible AI use, including protecting personal data.

The series concluded with reflections on what the students had learned, sparking curiosity about technology and encouraging a responsible approach to AI.

In the new episode of #digdeep, Christof Horn himself talks about the turnaround of the automotive industry.

Individual mobility is one of the freedoms that people value most. However, the mobility sector is undergoing unprecedented change.

Every two years, the IAA in Munich becomes the central meeting place for manufacturers, suppliers, and partners. In addition to the trade fair, automakers also showcase their predominantly electric models in the city center, while other providers of everything from public transport to innovative shuttles are represented.

In the podcast, Christof, as an industry expert, will discuss the transformation toward more software and connectivity, as well as the question of whether mobility is really improving.

Just listen in (podcast in German).

At this year’s DLD Future Hub in Munich, our PI Björn Ommer took the stage as one of the featured speakers.

In his keynote, he explored the transformative role of generative AI, emphasizing how advances in this field are set to reshape research, industry, and society at large.

At this year’s DLD Future Hub in Munich, our PI Fabian Theis delivered a keynote titled “Reinventing Science with AI”. He highlighted how AI is transforming research methodologies and driving innovation across science and industry.

Imagine trying to identify the full shape of a familiar object, e.g. a mug, when only its handle is visible. That’s the challenge a computer faces when estimating the pose of an object (its orientation and size) from partial data.

GCE‑Pose, a new approach from MCML Junior Members Weihang Li, Junwen Huang, Hyunjun Jung, Benjamin Busam, MCML PI Nassir Navab and collaborators Hongli XU and Peter KT Yu, introduced at CVPR 2025, tackles this with remarkable finesse.

This summer, I had the incredible opportunity to spend 11 weeks at Harvard University as part of the AI X-change program, visiting the group of Flavio Calmon at the Harvard John A. Paulson School of Engineering and Applied Sciences. The idea for this visit came after my former office mate and collaborator, Claudio Mayrink Verdun, a former member of MCML and PhD at TUM, moved to Harvard as a PostDoc. Since we had some ideas for new projects, and after meeting Flavio at NeurIPS last December, the plan to visit his group came together. I am very grateful to the AI X-change program for partially sponsoring this trip and making it possible, to Flavio for the kind invitation, and to Claudio for his support and collaboration during my stay.

What if robots could see in the dark and react faster than any human? Yannick Burkhardt, researcher at TUM and MCML, researches event cameras that capture motion in a revolutionary way. Unlike traditional cameras that take full pictures at fixed intervals, event cameras detect every tiny change in light—up to a million times per second—and feed this data into machine learning models.

MCML PI Niki Kilbertus has been awarded a prestigious ERC Starting Grant by the European Research Council.

The grant supports his project DYNAMICAUS, which combines machine learning, causal inference, and mechanistic modeling to study interventions in complex systems. By advancing methods that capture cause–effect relationships, the project opens new opportunities for understanding scientific processes and decision-making in real-world contexts.

ERC Starting Grants are among the most competitive European funding instruments, supporting excellent early-career researchers in building independent teams and groundbreaking projects.

Congratulations to Niki!

Can AI understand the world in 3D the way we do? Riccardo Marin, researcher at TUM and MCML, works at the intersection of computer vision and 3D geometry to teach machines how to perceive shapes, patterns, and spatial structures. His work ranges from detecting production flaws in manufacturing to analyzing archaeological artifacts – showing how understanding shape goes far beyond simple image recognition.

Our PI Gitta Kutyniok appeared as a guest on the IQ – Wissenschaft und Forschung podcast on ARD Audiothek. The episode, titled “Grüne Rechenzentren – Wie werden sie klimaneutral?”, explores how data centers can operate climate-neutrally despite rising energy demands from AI applications. Gitta Kutyniok joined other experts to discuss sustainable digital infrastructures and the role of AI in future computing.

Can AI help us understand why some people develop mental disorders while others remain resilient? MCML Junior Member Clara Vetter, PhD candidate in the group of our Director Daniel Rückert, uses machine learning to uncover hidden patterns in brain scans, genetic data, and even smartphone-based information. Her goal: identifying biological markers that could improve diagnosis and treatment in psychiatry.

We are proud to share that MCML Junior Member Shuo Chen and his team ranked among the Top 20 worldwide (out of 600+) in OpenAI’s Red-Teaming Challenge on Kaggle, a global competition that attracted over 5,900 participants.

Their work, titled “Bag of Tricks for Subverting Reasoning-Based Safety Guardrails”, investigates systemic weaknesses in reasoning-based safety mechanisms and provides insights into the development of more robust and template-independent alignment strategies.

Shuo Chen conducts his research in the group of our PI Volker Tresp.

Congrats from us!

Can AI from satellite imagery help us design more liveable cities, improve well-being, and ensure sustainable food production? Ivica Obadić, PhD student in the group of our PI Xiaoxiang Zhu, and MCML, develops transparent AI models that not only predict change but also give actionable insights for urban planners.

MCML PI Björn Ommer was a guest on the eighth episode of the IT-BUSINESS podcast IT ImPuls. Together with hosts Mihriban Dincel and Agnes Panjas, he discussed the current state of generative AI in Germany, the strategic opportunities it offers, and the importance of digital sovereignty and open-source initiatives for staying competitive in the global AI landscape.

Storm surges or collapsing dams: authentic simulations of water flows are not only important for special effects in disaster movies, but could also help to protect coastal regions. For more realistic simulations of fluid motions, MCML PI Nils Thuerey, Professor of Physics-based Simulation at TUM, and his team have developed a new method. Along with water, the method also takes into account the interaction with air. The approach is so efficient that calculations of complex wave motions can even be carried out with standard computers.

MCML Junior Member Mingyang Wang, PhD student in the group of our PI Hinrich Schütze, has been honored with the SAC Highlights Award at ACL 2025 for the paper “Lost in Multilinguality: Dissecting Cross-lingual Factual Inconsistency in Transformer Language Models.”

This award recognizes impactful research advancing the understanding of factual consistency across languages in large language models, contributing to more reliable and inclusive NLP systems.

Congratulations from us!

Azade Farshad, MCML Junior Member and PhD student in the research group of Nassir Navab, researches digital twins of patients at TUM and MCML to improve personalized treatment, surgical planning, and training. Using graph-based analysis and multimodal patient data, she builds models that create realistic surgical simulations — helping surgeons preview procedures, spot potential complications, and optimize strategies.

In the new episode of #digdeep, Philipp Baaske let us know, how he turned his doctoral research into the globally successful company NanoTemper, which offers innovative machines for the pharmaceutical industry.

The German education system opened up this opportunity for him, and he sees it as his duty to give something back to society. Starting in October, he will head the Entrepreneurship Initiative at LMU Munich, thus returning to his roots. Scaling research results differs in Europe from the US, but he isn't hesitant about pursuing other avenues besides multi-billion dollar valuations.

In his role, he wants to first listen, observe, and encourage students to become entrepreneurs.

Just listen in (podcast in German).

As a kid, Niki Kilbertus dreamed of becoming a theoretical physicist and discovering a fundamental law of nature. But when reality proved more complex, he found a new path through computer science.

Tracking actions, not just objects: The Spatiotemporal Action Grounding Challenge and Workshop at ICCV 2025 focusses on detecting and localizing actions both in space and time within complex, real-world videos.

Unlike standard action recognition, this task requires identifying when and where an action occurs, pushing models to handle long, diverse, and occlusion-heavy video content. Applications range from sports analytics to autonomous systems and large-scale video search.

The worksop is organized by MCML Junior Members Tanveer Hannan and Mark Weber, MCML Director Thomas Seidl and collaborators.

Imagine using social media to build a "living map" of our cities. Meet Shanshan Bai, MCML junior member and PhD student PhD student in the group of our PI Xiaoxiang Zhu, who's decoding the world around us with geo-tagged social media data.

Text-to-image (T2I) models like Stable Diffusion have become masters at turning prompts like “a happy man and a red car” into vivid, detailed images. But what if you want the man to look just a little older, or the car to appear slightly more luxurious without changing anything else? Until now, that level of subtle, subject-specific control was surprisingly hard.

A new method led by first-author Stefan Andreas Baumann, developed in collaboration with MCML Junior Member Felix Krause, co-authors Michael Neumayr, Nick Stracke, and Melvin Sevi, as well as MCML Junior Member Vincent Tao Hu, and MCML PI Björn Ommer changes this by giving us smooth, precise dials for individual attributes, like age, mood, or color, applied to specific subjects in an image.

Sven Nyholm, Chair of the Ethics of AI at LMU Munich and PI at MCML, explores one of the most urgent questions in AI: how responsibility, agency, and credit shift when intelligent systems make decisions for us.

In the recent ARD documentary "Depression – wie kommt man wieder raus?", MCML PI Björn Schuller shares insights on the role of artificial intelligence in mental health care.

He discusses how AI systems can support the diagnosis and treatment of depression by analyzing voice, facial expressions, and behavior patterns. These technologies pave the way for AI-based therapeutic assistants that may one day offer scalable, low-threshold support to those affected.

Meet Dominik Bär, MCML junior member and PhD student at LMU exploring how AI can enhance the integrity of social media platforms.

Dominik's work goes beyond just detecting harmful content like hate speech and misinformation. He's developing AI that understands why content is posted and generates thoughtful, human-like responses - so-called counterspeech - in real-time.

MCML PI Stephan Günnemann was recently featured in a BR24 segment discussing the rise of autonomous AI agents. He clarified the key distinction between AI assistants like ChatGPT and true agents: autonomy in both planning and execution. While companies such as Softbank envision agents taking over tasks from programming to corporate decision-making, Günnemann emphasized a more measured view, stating that humans will likely remain the final decision-makers, supported by intelligent agent systems.

MCML PI Fabian Theis has been awarded the 2025 ISCB Innovator Award by the International Society for Computational Biology. The award honors leading scientists who have made progressive contributions to computational biology and opened new directions in the field.

Congratulations from us!

This week, the Munich Center for Machine Learning hosted the event „KI und Schule im Dialog – Forschung, Bildungspolitik und Unterrichtspraxis” (AI and school in dialogue - research, education policy and teaching practice).

In the new episode of #digdeep, Rasmus Rothe explains, how the company Merantix from Berlin combines AI startups, investments, and applications.

Europe has developed significant artificial intelligence methods and plays a central role in industrial applications. The fundamental AI models are increasingly becoming common knowledge and readily available. However, success lies in combining AI with deep expertise and industry understanding.

One of the most successful AI startup communities has developed, particularly in Berlin, at Merantix, which has grown to 1,500 employees in eight years. This conversation with co-founder Rasmus Rothe will provide insights into how to attract top experts, the importance of the business idea versus the team, and how the startup culture can be further strengthened.

Just listen in (podcast in German).

Machine‑learning models can be undermined before training even starts. By silently altering a small share of training labels - marking “spam” as “not‑spam,” for instance - an attacker can cut accuracy by double‑digit percentages.

The paper “Exact Certification of (Graph) Neural Networks Against Label Poisoning” by MCML Junior Member Lukas Gosch, PIs Stephan Günnemann and Debarghya Ghoshdastidar and collaborator Mahalakshmi Sabanayagam, introduces the first exact guarantees that a neural network will remain stable under a prescribed number of label flips. Although demonstrated on graph‑neural networks (GNNs), the method applies to any sufficiently wide neural network.

From dreaming of seeing the Earth from space to leading efforts to understand our planet using AI and satellite data to tackle urgent global challenges. Xiaoxiang Zhu, Chair Professor for Data Science in Earth Observation at TUM and PI at MCML, develops machine learning systems that analyze petabytes of satellite imagery. Her work focuses on extracting reliable geo-information from raw data, especially in places where data is scarce or misleading.

MCML Junior Member Yusuf Sale, PhD student in the group of our PI Eyke Hüllermeier, has been honored with one of the IJAR Young Researcher Awards at ISIPTA 2025.

This prestigious award, presented by the International Journal of Approximate Reasoning, recognizes early-career researchers for outstanding contributions in the field of imprecise probabilities and uncertainty quantification.

Congratulations from us!

We are proud to announce that Barbara Plank, MCML PI and Professor of AI & Computational Linguistics at LMU Munich, has received the 2025 Imminent Research Grant for her work on language data together with Siyao Logan Peng and Marie-Catherine de Marneffe.

Congratulations to Barbara and the team!

MCML Director Daniel Rückert and his team are developing AI technologies to improve diagnostic imaging and protect patient data. Their research includes federated learning approaches that allow models to learn from clinical data without sharing sensitive information, as well as privacy-enhancing techniques like added data noise.

Their methods are already being applied in MRI and CT systems, leading to shorter exam times and more accurate diagnostics. The work is a key step toward integrating trustworthy AI into daily clinical practice.

Curious how AI is revolutionizing the treatment of eye diseases? Learn more about what AI can do for ophthalmology with Lucie Huang, MCML junior member and PhD student in the research group of Martin Menten, who is developing AI for sharper and faster eye scans. This means earlier diagnoses and better treatments for critical eye conditions like diabetes-related damage.

Discover how AI research is shaping our world and delivering solutions across diverse real-world problems. This video is the start of a series which is part of the project KI Trans, an initiative in collaboration with Junge Tüftler and LMU, focused on equipping teachers with the essential skills to navigate and teach about AI in schools. The project is funded by the Federal Ministry of Research, Technology and Space as part of DATIpilot.

Conversations, new connections and lots of pizza and drinks at our MCML Stammtisch, yesterday. Our Stammtisch is a regular informal meetup for all members. It’s a relaxed setting where researchers come together to chat about current projects, exchange ideas, or simply unwind and talk about life beyond the lab.

On July 22nd, we visited SAP Labs in Garching and had an event of our series “MCML Pitchtalks with Industry”.

Imagine looking at a single image and trying to understand the entire 3D scenery–not just what’s visible, but also what’s occluded. Humans do this effortlessly: when we see a photo of a tree, we intuitively grasp its 3D structure and semantic meaning. We learn this ability through interaction and movement in the 3D world, without explicit supervision. Inspired by this natural capability, our Junior Members–Christoph Reich and Felix Wimbauer–together with MCML Director Daniel Cremers and collaborators Aleksandar Jevtić, Oliver Hahn, Christian Rupprecht, and Stefan Roth from TUM, TU Darmstadt, and the University of Oxford, developed a novel approach: 🦖 Feed-Forward SceneDINO for Unsupervised Semantic Scene Completion. SceneDINO can infer both 3D geometry and semantics from a single image–entirely without labeled training data.

Often a new machine learning method claims to outperform the last. Whether it’s in bioinformatics, finance, or image recognition, the message is the same: this algorithm is faster, more accurate, more powerful. But can we trust those claims?

The stupid thing about artificial intelligence is that it is not intelligent. For any kind of intelligence, that is a hard saying to understand. Either the sentence is tautologous – if machine intelligence is not intelligent, it is stupid – or it sounds like an oxymoron, a contradiction in terms: If it is stupid, why do we refer to it as intelligence? What worries people who have dealings with AI is that both statements are correct: Artificial intelligence is at once glaringly stupid and breathtakingly intelligent.

We are happy to share that Eyke Hüllermeier, MCML PI will serve as spokesperson of the newly funded Research Training Group METEOR.

In the new episode of #digdeep, our PI Albrecht Schmidt talks about the possibility and chances of AI twins in the future.

Everyone uses AI based chatbot models. Would it also be possible to have our own AI model trained with our life data?

Such personalized AI models could act as digital twins, coaches, or advisors, thus expanding our capabilities. In our daily lives, we already collect a wealth of data about ourselves that could be used to develop such models. Although this idea may seem dystopian at first glance, at second glance it offers a fascinating opportunity for more democratic and individualized support.

Prof. Albrecht Schmidt from LMU Munich focuses on the collaboration between humans and computers and encourages reflection on the future digital world.

An exciting excursion into the question of what kind of digital world we want and can live in in the future!

Just listen in (podcast in German).

We are proud to share that the paper “The Value of Prediction in Identifying the Worst-Off” by our Junior Member Unai Fischer Abaigar, and Associate Christoph Kern, and collaborator Juan Carlos Perdomo from Harvard University has been selected for an Outstanding Paper Award at ICML 2025.

Only six papers received this prestigious recognition this year, highlighting the exceptional quality and impact of their work.

Congrats from us!

Christian Wachinger, Professor of AI in Radiology at the Technical University of Munich and PI at MCML, is developing AI systems to precisely reconstruct the brain’s cortex — an extremely thin and highly folded sheet of neural tissue — in order to measure it for diagnostic use.

MCML PI Fabian Theis has been awarded a prestigious ERC Advanced Grant by the European Research Council.

The grant supports his project “DeepCell”, which aims to develop advanced deep learning methods for analyzing dynamic single-cell data. By combining state-of-the-art machine learning with cutting-edge biological research, the project seeks to better understand cellular behavior over time, a key step toward improving diagnostics and personalized medicine.

ERC Proof of Concept Grants are highly competitive funding instruments awarded exclusively to previous ERC grantees, enabling them to bridge the gap between frontier research and potential innovation or commercialization.

Congrats from us!

Our PI Nassir Navab has received the 2025 IEEE EMBS Technical Achievement Award.

He was recognized “for pioneering roles in the establishment of the medical augmented reality, surgical data science, and robotic imaging fields.”

The award, presented annually by the IEEE Engineering in Medicine and Biology Society (EMBS), honors individuals for outstanding achievements and innovations in the field of biomedical engineering.

Congrats from us!

In May, a 20-member delegation from MCML visited top U.S. institutions — Harvard, MIT, NYU, and Cornell Tech — to foster long-term collaboration in AI research. The visit marked an important step in MCML’s newly launched AI X-Change program, which supports international research stays for PhD students and hosts visiting scholars in machine learning. The trip focused on research in generative models and medical AI—covering topics such as fairness, transparency, clinical applications, and the integration of AI into healthcare. Discussions also explored the foundations of machine learning and social data science.

Imagine an operating room - a space filled with intricate interactions, rapid decisions, and precise movements. Now, imagine capturing every detail of such a complex environment not just visually but also through sound, dialogue, robot movements, and much more. This is exactly what MM-OR -- A Large Multimodal Operating Room Dataset for Semantic Understanding of High-Intensity Surgical Environments accomplishes.

"A good prediction is rarely a good decision." This may sound surprising in an age where AI is often celebrated for its predictive power. But when it comes to making real-world decisions — especially in complex environments like business or healthcare — prediction alone doesn’t cut it. What we need is AI that can help us choose what to do, not just tell us what might happen.

At the MCML Munich AI Day, we held a Masterclass on Entrepreneurship, with two invited speakers from our collaboration partners, Annie Weichselbaum from the LMU Innovation & Entrepreneurship Center, and Bastian Burger from TUM Venture Labs. Both gave short impulses on how to build a company, how to pitch, on what matters when you try to build your start-up, finding investors, learning to accept failure, and to be resilient. Thank you so much for this collaboration - Bastian and Annie!

The MCML Munich AI Day 2025 began on a collaborative note with a masterclass on science communication, in collaboration with the RHET AI Center and the Tübingen AI Center. A group of PhD researchers and PostDocs came together to tackle a crucial challenge: how to make their AI research more accessible, engaging, and impactful beyond academic circles.

The MCML Munich AI Day 2025 brought together a full house of leading researchers, policymakers, industry experts, and journalists to discuss the transformative power of AI in today’s society. In addition,as part of the ELLIS Pre-ICML 2025 Fest, the event served as a vibrant platform for interdisciplinary dialogue on the societal, scientific, and technological implications of AI.

MCML Principal Investigator Anne-Laure Boulesteix has been awarded a prestigious Reinhart Koselleck Grant by the German Research Foundation (DFG). The highly competitive program supports researchers with an outstanding scientific track record who pursue particularly innovative or high-risk projects.

Her project, titled “The design, interpretation and reporting of empirical studies evaluating statistical methods”, focuses on how studies that test statistical methods are planned, analyzed, and reported. The project is expected to connect well with MCML’s work on Open Science, EmpiricalML, and metascience.

Congrats from us!

Clear imaging is essential for accurate diagnosis — but when it comes to the heart, motion makes it one of the most difficult organs to capture in high resolution. Traditional reconstruction methods often struggle to deliver the detail clinicians need.

We are pleased to share a video message from of the Federal Ministry of Research, Technology and Space (BMFTR) on the occasion of the MCML Munich AI Day 2025.

In her address, PSt’in Dr. Silke Launert emphasizes the federal government’s strong commitment to artificial intelligence as a key technology for the future. She highlights the importance of responsible, application-oriented AI research and the role of transfer research in bringing scientific findings into practice.

Dr. Launert commends the Munich Center for Machine Learning for its contribution to Germany’s AI ecosystem and its efforts in strengthening collaboration between academia, industry, and policy. She also underlines the value of events like the MCML Munich AI Day in building international networks and encouraging dialogue on ethical, innovative AI.

We thank the BMFTR for its continued support and for sharing this encouraging message with the MCML community.

Bluetooth toothbrushes, deeply embedded software, treacherous doorbells: At the heart of the International Conference on Applied Cryptography and Network Security (ACNS), which is taking place from 23 to 26 June, is the question of how we can make our increasingly connected everyday world more secure. Co-organizer and MCML Associate Johannes Kinder, Chair Professor of Programming Languages and Artificial Intelligence at LMU, explains how ordinary people can secure their households, what manufacturers need to do – and why lawyers and ethnographers are researching digital security.

We are happy to share that our PI Zeynep Akata has been awarded the 2025 ZukunftsWissen Prize by the German National Academy of Sciences Leopoldina and the Commerzbank Foundation.

The award honors her work in explainable and trustworthy AI, including early contributions to zero-shot learning and recent advancements in multimodal generative models. Akata’s research focuses on making AI decisions transparent and interpretable, helping build systems that explain their outputs in human-understandable ways.

The prize, endowed with €50,000, will be officially presented during the Leopoldina Annual Assembly on September 25, 2025, which this year focuses on artificial intelligence.

Congratulations from us!

Our PI Volker Tresp was recently interviewed by Plattform Lernende Systeme – Germany's AI Platform. In the interview, he discusses Germany’s position in the international AI race, the challenges AI start-ups face, and the lessons that can be learned from the U.S. innovation ecosystem. As head of the working group “Technological Enablers and Data Science”, Tresp offers valuable insights into how Germany can foster cutting-edge AI research and build competitive AI infrastructure.

In presidential debates, every word counts - not only what candidates say but how they say it. Words are carefully selected to resonate with voters' values, fears, and hopes. A new insightful study by MCML Junior Member Philipp Wicke and Marianna M. Bolognesi dives deep into the September 10, 2024 debate between Donald Trump and Kamala Harris.

Artificial intelligence has enormous potential in radiology — but realizing it requires more than good algorithms.

Michael Ingrisch, Clinical Data Science Professor at LMU and MCML PI, shares how his team took a practical approach: identifying a key diagnostic challenge in PET CT imaging and inviting the broader AI community to help solve it.

How can we guarantee that autonomous vehicles always make the right decision in unpredictable traffic?
Matthias Althoff, Professor at the chair of Cyber-Physical Systems at the Technical University of Munich and MCML PI, explores the complex challenge of ensuring safety in AI-powered driving systems.

MCML PI Massimo Fornasier has been awarded an ERC Advanced Grant for his project NEITALG. The project aims to develop new algorithms that can reliably find global solutions to challenging nonconvex optimization problems, with applications in drug development and AI safety.

Congrats from us!

Imagine watching a two-hour video and trying to find the exact moment someone scores a goal - or says something important. Humans can do this with ease by skimming and zooming in. But for AI, finding specific moments in long videos is incredibly hard.

Most current AI systems struggle to handle more than a few minutes of video at a time. They often miss the forest for the trees - either losing detail or skipping over important moments. That's where ReVisionLLM comes in.

ReVisionLLM, developed by MCML Junior Member Tanveer Hannan, MCML Director Thomas Seidl and collaborators Md Mohaiminul Islam, Jindong Gu and Gedas Bertasius, mimics how humans search through content: by starting broad, identifying interesting segments, and then zooming in recursively to locate precise start and end points of an event.

MCML PI Albrecht Schmidt has been awarded an Advanced Grant by the European Research Council (ERC) for his project “AI of Human Experience: Towards Personal Generative AI‑Systems for Amplifying Human Cognition.” The project explores the development of personalized generative AI systems designed to support human memory, planning, and creativity.

Congrats from us!

MCML Junior Members Tong Liu, Gengyuan Zhang, Shuo Chen, and MCML PI Volker Tresp and their co-authors have received the Best Paper Award at the Second Workshop on Responsible Generative AI (ReGenAI) Workshop at CVPR 2025 for their paper “Multimodal Pragmatic Jailbreak on Text-to-image Models”.

The authors show that text-to-image models can be easily exploited to produce unsafe content through cross‑modal interactions between safe text and images, a vulnerability that current safety filters fail to address effectively.

Congratulations from us!

As algorithms increasingly make decisions that impact our lives, from managing city traffic to recommending hospital treatments, one question becomes urgent: Can we trust them?

How does data quality shape the future of AI? Frauke Kreuter, Professor at the Chair of Statistics and Data Science at LMU Munich and MCML PI, shares her insights on the fundamental role of data quality in the development and deployment of AI.

At re:publica 2025 in Berlin, our PI Björn Ommer gave the opening keynote on how generative AI is transforming intelligence into accessible, scalable knowledge. This shift moves society from raw information to decision-relevant knowledge, empowering individuals.

Björn highlighted how personal generative AI will change our thinking and decision-making, and emphasized the importance of a mindset shift to ensure European AI sovereignty and benefit for all.

MCML Junior Research Group Leader Benjamin Lange was recently featured on the mCAST podcast. He talks about the ethical challenges of using artificial intelligence in mobility, explains the concept of Human-in-the-Loop, and discusses how AI is already applied today and what future possibilities it holds.

MCML Director Daniel Rückert has been elected to the Fellowship of the Royal Society, UK’s national academy of sciences, in recognition of his contributions to artificial intelligence in medicine and healthcare.

The Fellowship is awarded to individuals who have significantly advanced knowledge in their field.

The Fellows join the ranks of Stephen Hawking, Isaac Newton, Charles Darwin, Albert Einstein, Lise Meitner, Subrahmanyan Chandrasekhar and Dorothy Hodgkin.

Congrats from us!

Recent advances in Natural Language Processing (NLP) and Large Language Models (LLMs) are transforming how we engage with and interpret human-generated data—highlighting critical issues around data quality, bias, and ethics. At the same time, Public Opinion Research (POR) offers well-established methods for collecting high-quality, representative data from human subjects.

This interdisciplinary workshop, co-organized by MCML members, brings together experts from both fields to explore how POR can inform and strengthen NLP research, and how NLP methods can enhance POR. The goal is to foster dialogue and collaboration around data quality, methodological transparency, and ethical AI development in the age of LLMs.

Deadline for paper submission: June 23, 2025

The workshop is co-organized by our PIs Frauke Kreuter and Barbara Plank.

Visit the website of NLPOR for more information!

We visited Daiichi Sankyo at their Munich Office and learned a lot about their great work in Pharma and AI/ML.

In the next episode of #digdeep, Christoph Keplinger gives information about advances in robotics.

Even though humanoid robots are increasingly adopting human geometries, with their hard metal forms, they remain much less flexible than humans with adaptable muscles and highly sensitive grippers. Artificial muscles could bring many advantages to robots and also make collaboration between humans and robots safer.

Our studio guest, Prof. Christoph Keplinger, heads the Max Planck Institute for Intelligent Systems (MPI-IS) in Stuttgart after working at Harvard and Colorado. We discuss advances in robotics, particularly the transition from hard to soft materials and the mimicking of natural muscles.

We explore the importance of embodied intelligence and the future of wearable robotics that can be integrated into everyday life. What would it be like if artificial muscles were integrated directly into our clothing?

We want to hear from him about the role creativity plays in the design process and the challenges that arise when translating research into practice. Christoph also shares his experiences as a startup founder and the lessons he's learned from his mistakes.

Just listen in (podcast in German).

MCML Junior Member Christian Koke, PhD student in the group of our Director Daniel Cremers, and his co-authors have received the ICLR 2025 MLMP Best Paper Award for their paper “On Incorporating Scale into Graph Networks”. The award includes 2,000 GPU-hours from Nebius and honors their outstanding contribution to multiscale graph network research.

Congratulations from us!

Hate speech on social media isn’t just offensive - it’s dangerous. It spreads quickly, harms mental health, and can even contribute to real-world violence. While many studies have focused on identifying hate speech or profiling those who create it, a key piece of the puzzle remained missing: Who reshares hate speech?

The team at MCML - Dominique Geissler, Abdurahman Maarouf, and PI Stefan Feuerriegel - explored this question with their latest work: "Analyzing User Characteristics of Hate Speech Spreaders on Social Media".

MCML Junior Member Emanuel Sommer is a PhD-student at the Munich Uncertainty Quantification AI Lab at LMU Munich supervised by our PI David Rügamer. His research focuses on Scalable and Reliable (Bayesian) Deep Learning.

Generative AI and medical AI are at the forefront of technological innovation, offering transformative potential across sectors. A core delegation of around 15 researchers from MCML will showcase and discuss research in Generative AI and Medical AI, aiming to establish new partnerships and collaborative research initiatives.

The MCML visited Bain & Company for another event of our series "MCML Pitchtalks with Industry".

Our PI Gitta Kutyniok has been invited to join the National Academy of Artificial Intelligence (NAAI). The invitation acknowledges her contributions to harmonic analysis, compressed sensing, and AI, as well as her recognition as SIAM Fellow (2019) and IEEE Fellow (2024).

Congratulations from all of us at MCML!

Imagine standing in an unfamiliar part of a city, no GPS in sight. All you can say is, "I’m west of a green building, near a black garage." That might be vague to a machine, but Text2Loc understands you perfectly. With this powerful new system, AI can find your exact location in a 3D map - just from how you describe the world around you.

As part of this year’s Girls’ Day, we were excited to welcome a group of curious and motivated students to our workshop: “Learn how AI works and impacts our world.”

We hosted another edition of the MCML Pitchtalks with Industry – this time in collaboration with appliedAI and their partner Atruvia AG.

The MCML was honored by the visit of Luc Vinet, Foutse Khomh, and Josée Savard from IVADO, Montreal. IVADO is a research consortium led by Université de Montréal to advance robust, reasoning, and responsible AI.

We talked about transatlantic scientific cooperation between Germany and Canada and also about the importance of research stays for young scientists. The MCML presented its newly established AI X-Change program, in which we offer our junior members support for research stays abroad.

In another new episode of #digdeep, Philipp Noll, Co-Founder from Spread.AI talks about data networking in large-scale IT projects.

Data is of central importance in modern industry, both in product engineering and in the networking of the entire process chain, which ranges from research and development to production, sales, and after-sales. Despite its importance, data is often viewed as a byproduct, which leads to inadequate organization: it is often scattered, incomplete, and lives in a multitude of disconnected IT systems.

The scale-up Spread.AI has developed innovative solutions to overcome these challenges by relying on intelligent networking through knowledge graphs and connectors.

In conversation with co-founder Philipp Noll, the importance of data as the "new gold" for engineering and the opportunities that arise from the networking of data worlds are highlighted. With a platform that enables the automation of tasks in the development process, the path is paved for the widespread use of AI agents in engineering.

Just listen in (podcast in German).

In the new episode of #digdeep, the podcast reports about circular economy at Schwedt and a startup-lab with Sascha Lademann.

Schwedt, a town in the Uckermark region, is facing the challenge of structural change, as its traditional industries, particularly oil refineries, are no longer viable. The region is seeking new ways to shape its economic future, relying on the principles of the circular economy to close material cycles. Furthermore, the energy supply is being optimized through innovative concepts such as Power2X, which intelligently connect producers and consumers.

To support the transition to a climate-neutral economy, renewable energies and new infrastructure projects are being promoted, supported by extensive funding programs from the federal, state, and EU governments. Digitalization plays a crucial role in this transformation process, which is why a startup lab was established to make the necessary data and processes transparent.

So, Schwedt should be everywhere!

Just listen in (podcast in German).

Understanding the location and semantics of objects in a scene is a significant task, enabling robots to navigate through complex environments or facilitating autonomous driving. Recent AI models for understanding scenes from images require significant guidance from humans in the form of pixel-level annotations to achieve accurate predictions.

To overcome the reliance on human guidance, our Junior Members - Christoph Reich and Nikita Araslanov - together with MCML Director Daniel Cremers and collaborators Oliver Hahn, Christian Rupprecht, and Stefan Roth from TU Darmstadt and the University of Oxford, proposed a novel approach: 🥤🥤 CUPS: Scene-Centric Unsupervised Panoptic Segmentation.

“I spend a lot of time explaining to my students, but also to managers from a wide variety of businesses, all the things that AI cannot do”, MCML PI Stefan Feuerriegel says. His job consists to a large extent in “dialing back the excessive expectations. When I’ve accomplished that, then we can start to work successfully with AI.”

From March 28–30, LMU Munich hosted DataFest Germany 2025, bringing together 24 student teams from 15 universities for an intense 48-hour data challenge. Participants worked with a newly developed dataset on German municipalities, tackling real-world questions through data analysis and storytelling.

A research team at TUM, led by our PI Angela Schoellig, has developed the mobile robot Jack, which moves safely and efficiently through crowds.

This is made possible by an innovative motion planning algorithm inspired by human behavior. Jack continuously scans its surroundings using a lidar system and updates its trajectory ten times per second, enabling smooth, collision-free movement—even in dynamic environments.

The technology offers promising applications in service robotics and autonomous vehicles, where natural and safe interaction with people is essential.

Machine learning models make powerful predictions, but can we really trust them if we don't understand how they work? Global feature importance methods help us discover which factors really matter - but choosing the wrong method can lead to misleading conclusions. To see why this is important, consider a real-world example from medicine.

Our Junior Member Azade Farshad was awarded the BVM Award for the best thesis in medical image processing. Her work, supervised by Nassir Navab, was recognized for its innovation and excellence.

The award was presented at the German Conference on Medical Image Computing (BVM), held at OTH Regensburg.

Congratulations from all of us!

More than 170 audiences visited the online lecture of our Junior Member Ruotong Liao, PhD student in the group of our PI Volker Tresp, on Monday, 17. March 2025, as an invited speaker at the United Nations "AI for Good".

With her talk "Perceive, Remember, and Predict: Explainable Multimodal Agents with Symbolic Representations," Ruotong Liao took part in the online event "Explainable Multimodal Agents with Symbolic Representations & Can AI be less biased?"

At the event, which was hosted by the leading platform for artificial intelligence for sustainable development, Ruotong Liao explained her research results, focussed on how the integration of temporal reasoning and symbolic knowledge about evolving events enables LLMs to make structured, interpretable, and context-sensitive predictions. Ruotong Liao presented work aimed at developing explainable multimodal agents capable of perceiving, storing, predicting, and justifying their conclusions over time.

See the whole presentation in the stream.

Researchers at the Technical University of Munich have developed an innovative method that drastically lowers the energy consumption of artificial intelligence systems. The approach optimizes computational efficiency, making AI applications more sustainable and cost-effective.

Our Associate Felix Dietrich emphasized the importance of energy-efficient AI, highlighting its potential to reduce environmental impact while maintaining high-performance capabilities.

Despite their impressive capabilities, Text-to-Image (T2I) models frequently misinterpret detailed prompts, leading to errors in object positioning, attribute accuracy, and color fidelity. Traditional improvements rely on extensive dataset training, which is not only computationally expensive but also may not generalize well to unseen prompts. To address this, our Junior Members – Luca Eyring, Shyamgopal Karthik, and Karsten Roth – together with PI Zeynep Akata and collaborator Alexey Dosovitskiy from inceptive, proposed a new approach: Reward-based Noise Optimization (ReNO).

Our PI Björn Schuller was featured in the new ARTE TV documentary on "Therapy via AI". In the documentary, he briefly outlined the focus of his research. The AI program he presented is capable of recognizing the emotion behind spoken words, demonstrating how artificial intelligence can analyze vocal expressions to assess emotional states. Together with his team, he has also developed an app that could potentially be used as a virtual therapist in the future.

During my research stay at Broad Institute of MIT and Harvard in autumn 2024, I had the pleasure of being part of the research group led by Caroline Uhler, Director of the Eric and Wendy Schmidt Center (EWCS) at the Broad Institute, and Andrew (1956) and Erna Viterbi Professor of Engineering in EECS and IDSS at MIT. My three-month stay in Boston gave me the opportunity to investigate the application of causality and dynamical systems in biology. This research experience was made possible through the generous support of the MCML AI X-change program.

Medical reports, especially in radiology, are commonly difficult for patients to understand. Filled with complex terminology and specialized jargon, these reports are primarily written for medical professionals, often leaving patients struggling to make sense of their diagnoses. But what if AI could help? Recognizing this potential early on, the team of our PI Michael Ingrisch launched a study just four weeks after ChatGPT’s release to explore how it could simplify radiology reports to make them more accessible. The study has now also been published in European Radiology.

In a recent interview with Plattform Lernende Systeme, our PI Gitta Kutyniok shared her insights on the current AI developments. She discussed what Europe needs to stay competitive, how research can better translate into real-world applications, and why mathematics plays a key role in making AI more explainable.

In the new episode of #digdeep, the podcast explains reasons for digital optimism.

Many people have ideas - but how do you find a really good business idea and how do you turn the idea into a successful business model?

Dr. Alex Mrozek is not only the CEO of Oetker Digital, but with his platform DigitaleOptimisten.de he brings future founders, ideas and investors together. And with his very successful podcast he is also active in the Brand Eins family. And each of his studio guests has to bring a business idea with them - sharing instead of hoarding is the motto.

Technological and social change has never been as fast as it is today. But there have never been so many opportunities to implement your own ideas and skills. So it's time for us to get infected by his digital optimism!

Just listen in (podcast in German).

Our PI Angela Schoellig, was a guest on the Blaue Couch podcast from Bayerischer Rundfunk, where she talked about the importance of mathematics in everyday life and its role in robotics. She shared insights into her research and how math helps robots make smart decisions in real time.

MCML Director Daniel Cremers met the French Minister of Science Philippe Baptiste at the AI Action Summit in Paris, last week.

He was also part of the Symposium about frontiers in Generative AI at the Institut Polytechnique de Paris that hosted the scientific days of the AI Action Summit. Together with Vicky Kalogeiton, Gaël Varoquaux, Stéphane Mallat, and Lingpeng Kong he explored the mathematical foundations, scalability solutions, and future directions of GenAI, emphasizing the role in AI-driven scientific discovery, creative applications, and computational science’s broader landscape.

OpenAI is set to open its first office in Germany, recognizing Munich’s strengths as a hub for innovation.

Reinhard Heckel, MCML PI, highlighted Munich's deeptech scene, its talented workforce, and its strong presence in industries like biotech and venture capital. "Our infrastructure here is a bit like Silicon Valley," Heckel noted in an interview with BR24.

On February 10, 2025, MCML JRG Leader Benjamin Lange appeared on the ZDF program heute journal to discuss the impact of Artificial Intelligence on higher education.

He pointed out that writing is an important part of the thinking process and shouldn’t be fully automated, especially in the humanities.

With AI becoming more present in academia, universities are now figuring out how to adapt exams and keep the core of scientific research intact.

MCML PI Gitta Kutyniok and her team at LMU are assisting companies in tackling the challenges of the EU’s AI Act.

Their research focuses on creating practical solutions to ensure compliance, while also supporting innovation. With AI regulations constantly evolving, their expertise is crucial in bridging the gap between policy and technology.

Our Junior Member Felix Köhler, together with our PIs Rüdiger Westermann and Nils Thuerey, and collaborator Simon Niedermayr, have introduced APEBench, an innovative benchmark suite. APEBench sets a new standard for evaluating autoregressive neural emulators, which are designed to solve partial differential equations (PDEs)—the fundamental mathematical framework for modeling natural phenomena such as weather patterns, fluid dynamics, and chemical reactions.

In the February episode of #digdeep, the podcast highlights DeepSeek, OpenAI & Co.

When the supposed startup DeepSeek from China presented a competitive model at supposedly the lowest cost, the plate tectonics of AI conditions seemed to be breaking down: Who is leading the market, how far ahead do the already established players have, and do we really need as much energy and resources as we thought?

The violent reactions in the press (and on the stock market) perhaps show above all how important the major foundational models have become for our economy and society - and that the massive investments are mainly intended to keep others off the field.

So it's high time for a studio guest who really knows his stuff about AI! Prof. Kristian Kersting heads the Machine Learning/AI Lab at TU Darmstadt, was voted one of the top 100 most influential AI researchers, and the list of his awards and memberships in the AI ​​community would go beyond the scope of the podcast. The best conditions to ask: What is actually happening in AI right now: revolution or little revolution? And when will the singularity finally arrive, when the machines will overtake humanity?

You can find out whether there is already an answer to this in our very entertaining conversation with human intelligence!

Just listen in (podcast in German).

MCML Junior Member Gabriel Marques Tavares has a PostDoc position at the chair of Database Systems and Data Mining of our Director Thomas Seidl. His research focus is in the field of Process Mining, which investigates the execution of business processes within organizations, aiming at improving their performance, saving resources and identifying bottlenecks.

Every year, academics, the leading online career portal for science, research, public affairs and society, honours a person who has made a lasting contribution to the respective scientific field with outstanding and forward-looking research achievements and who has also distinguished themselves through exemplary action and voluntary commitment to science.

MCML Associate Steffen Schneider and his team are developing machine learning algorithms to statistically model dynamic processes in biological systems. In this way, they are creating new approaches for basic medical research. He also demonstrates a remarkable social commitment: with his initiative ‘KI macht Schule’ (AI goes to school), he has been bringing the topic of artificial intelligence into the classroom since 2019.

Congrats from us!

In the :newstime segment from January 28, 2025, MCML PI Reinhard Heckel commented on the new AI platform DeepSeek, highlighting its potential to enable the creation of high-quality models at low cost.

The Bavarian State Center for Political Education again hosted “Lernort Staatregierung", a program at the Bavarian State Ministries and the State Chancellery in Munich, as well as at the offices of the Ministries of Finance and Health in Nuremberg.

The program offers young people a first-hand, interactive look at the structures, responsibilities, and daily work of the Bavarian state government. Through direct interaction with decision-makers in politics and public administration, students gain a tangible understanding of how political decisions are made.

During the visit to the Bavarian State Ministry for Science and the Arts (StMWK), the MCML presented a lecture on AI. The session provided students an accessible introduction to current AI research and the social and political implications of these technologies. The talk was led by MCML Junior Members.

MCML Junior Member Dominique Geissler, PhD student in the group of our PI Stefan Feuerriegel, was recently featured in Stern, where she spoke about the growing concerns surrounding AI-generated deepfakes in the political sphere.

The host of alpha-gespräche, Mirjam Kottmann, discusses the possibilities that AI opens up for healthcare and healing with MCML PI Fabian Theis and Daniela Hartmann, Chief Physician at the Clinic for Dermatology, Allergology and Laser Medicine at the München Klinik.

Researchers from the Technical University of Munich (TUM) and Helmholtz Munich have tested self-supervised learning as a promising approach for analysing over 20 million individual cells. MCML PI Fabian Theis holds the Chair of Mathematical Modelling of Biological Systems at TUM. Together with his team, he has investigated whether self-supervised learning is better suited to analysing large amounts of data than other methods.

Given the enormous amounts of data generated by advances in single-cell technology, it is important to interpret them efficiently in order to recognise differences between healthy cells and those with diseases such as lung cancer or COVID-19. Self-supervised learning does not require pre-classified data and enables the robust processing of large amounts of data.

The study shows that this method improves performance, especially on transfer tasks and zero-shot predictions. The researchers compare masked learning and contrastive learning and find that masked learning is better suited for large datasets. The results could lead to the development of virtual cells that map the diversity of cells in different datasets and are useful for analysing cell changes in diseases.

We are proud to announce that our PI Björn Ommer was a featured speaker at DLD Munich 2025, celebrating 20 years of shaping the future.

Björn delivered an inspiring talk titled "Generative AI: Future-proofing Germany’s Innovation Hub", exploring the transformative potential of generative AI and its role in maintaining Germany's leadership in innovation.

We are proud to share that our PI Gitta Kutyniok was an invited speaker at the prestigious DLD Conference 2025, participating in the BAIOSPHERE track.

In her talk, titled "Next Generation AI Computing", Gitta highlighted the critical role of advanced computing in ensuring the sustainability and reliability of future AI systems. She addressed challenges such as energy consumption, reliability, and compliance with the EU AI Act, while also exploring solutions like analog, neuromorphic, quantum, and biocomputing.

This talk also introduced the GAIn project, a collaborative effort with colleagues from TUM and TU Dresden, aimed at developing energy-efficient and trustworthy AI systems.

In the new episode of #digdeep, the podcast makers visited the Work Awesome Conference in Berlin.

At a time characterized by multiple crises - such as the war of aggression in Ukraine, the Hamas attack on Israel and the escalation in the Middle East conflict - companies are facing major challenges. These issues not only lead to intense discussions in society, but also to conflicts in the private and professional sphere.

At the Work Awesome conference in Berlin, organized by Brand Eins, #digdeep addressed the question of how companies deal with these challenges and what role their values play in this. Where does an employee's identity as part of the organization end and where does their private self begin? The guests, Jörg Ehmer, CEO of Burger King Germany, and Judith Muster, sociologist and partner at the management consultancy Metaplan, provided fascinating insights. The discussion was moderated by Frank Dahlmann, who heads the B1 Podcast Network. It was an enriching premiere that offered important perspectives on current social challenges.

Just listen in (podcast in German).

MCML PI Alena Buyx was featured in the ARD Wissen documentary “Mein Mann lebt als KI weiter”, discussing the ethical implications of AI technologies that replicate human behavior. She explored moral questions arising when AI creates digital versions of deceased individuals or AI-based partners, highlighting how these innovations challenge our understanding of intimacy, love, and mortality.

We are happy to share that Junior Member Sarah Ball, a PhD student in the group of our PI Frauke Kreuter, has had her research featured on the LLM Cybersecurity podcast by Ariel Fogel and Tyler Bettilyon.

The discussion centered around Sarah's paper "Understanding Jailbreak Success: A Study of Latent Space Dynamics in Large Language Models" highlighting its insights and implications for the field of AI and cybersecurity.

In their recent work, "Liar, Liar, Logical Mire: A Benchmark for Suppositional Reasoning in Large Language Models" our Junior Member Philipp Mondorf and our PI Barbara Plank tackle a fascinating question: How well do AI systems handle complex reasoning tasks?

In the Science Date segment on NANO (3sat), our PI Fabian Theis discusses key questions on AI and its role in healthcare: Can we recreate the human brain? Will AI make humans obsolete? And how is AI transforming medical research? He suggests that AI can make healthcare more accessible, affordable, and democratic, though these profound changes raise complex ethical and scientific considerations.

The first 2025 episode of #digdeep reflects on twelve exciting topics and interviews in 2024 that inspired both podcast creators and listeners.

Whether in mechanical engineering, where even medium-sized companies can get ahead with attractive digital concepts, or in medicine and healthcare, where new start-ups are now making research data usable in various areas: Only with the digital transformation, which must have arrived in Germany long ago, the two podcasters believe, will it become apparent what economic leaps are possible and how well data can help to steer and regulate issues in society. “They are quantitative accompanying research for social development,” says MCML PI Frauke Kreuter.

In evidence-based politics and data analysis, research is also continuously focused on bringing about a change of mindset in the digital sphere. While public opinion is important for election forecasts in election polls, reporting is also important in election campaigns. In the coming years, it will be exciting to see what role science and the press play in the political spectrum. Broadcasters and the press in general are working intensively on AI. “AI is becoming more and more prevalent, helping, supporting and taking away routine work, but it should be seen as a very powerful tool,” says Horn.

Artificial intelligence, which is determined by hype cycles and challenges, will enter the next productive phase in 2025. ChatGPT and generative AI show various parallels and partly shifted hype cycles. Depending on the task for which generative language models are being used. “Data analysis work can be extremely accelerated if parts of code are generated,” says MCML PI Frauke Kreuter. “It's fascinating to see what's happening right now. The experts are getting faster, the beginners are getting a jump start and everyone who doesn't use it is simply left behind in terms of speed.” When it comes to alignment, Kreuter is also concerned with how to get language models to do what you want or don't want them to do. “Guardrails and newer models are so powerful that you can hardly keep up with them in research,” says Kreuter, ”other models may be needed here in the future.”

For Christof Horn, his personal hype cycle is multi-layered. ChatGPT not only helps in the daily search for topics and delivers precise results, but also in understanding topics and acquiring world knowledge, for example with regard to language or technology.

In addition to the euphoria and democratization, the limits of the models, which cannot do any real reasoning, are now becoming clearly visible: “We are currently experiencing a crisis for everyone. The idea that even more training data will help does not work and there is a saturation of the models, because even more parameters do not lead anywhere. Moreover, the training data is not getting any better, but is contaminated with results that it has produced itself,” says Horn.

The AI topics on the hype cycle continue: “What we will see next year is an explosion in the integration of generative AI into data production,” says MCML PI Frauke Kreuter. The chatbot as an interviewer and the automatic evaluation of open answers. Both will become decisive criteria for the future of the customer journey.

To escape the perpetual feeling of inevitable chatbot hell in 2025, real people will gain in value. Engineering and other tools will nevertheless bring AI further among us and the discussion about language models will no longer be held. Robotaxis, lawn robots and digital agents that you can talk to will remain normal for people. Also because they are entertaining. When it comes to influencing, every medium offers a new platform.

In autonomous driving, platforms such as “Waymo”, which work with training data and further optimize it, will have a stronger influence, as will innovative, promising start-ups, such as those in Germany in sensor technology or international ones that automate delivery and factory traffic.

In their podcast, MCML PI Frauke Kreuter and Christof Horn 2024 also dealt with the digitalization of shoes and education platforms, were at the “herCAREER” career fair and were the first German-language podcast to host ChatGPT for the second time.

With or without a language model, but not yet generated by it: #digdeep will still exist in 2025. In the upcoming January episode with “BrandEins”, they visit the “work awesome” conference in Berlin with the CEO of “Burger King” and “Metaplan”.

Just listen in (podcast in German).

During the summer of 2024, I had the privilege of representing the Computer Vision Group at TUM, led by MCML Director Daniel Cremers, in a collaborative research project with the Geometric Computing Group at Stanford University, headed by Prof. Leonidas Guibas. During my three-month stay at Stanford, researchers from both institutes delved deeply into the fascinating intersection of geometry and machine learning, launching research projects initiatives aimed at advancing the edge of the current state of the art. The outcomes of our ongoing efforts will be submitted for publication in leading computer vision and graphics venues. This collaboration is generously funded by the MCML AI X-change program and BaCaTeC.

Our PI, Angela Dai, focuses on the application of artificial intelligence to design and optimize complex systems. Her research contributes to the development of AI technologies that enable faster and more efficient designs in the automotive industry. By using advanced algorithms, her work helps reduce development costs and enhance the sustainability of products.

Thank you for a successful year of collaborations and partnership! Cheers to a festive season!

We look forward to seeing you in 2025!

The Open Science Initiative in Statistics - which is part of the Open Science Center at LMU Munich - and MCML recently hosted a workshop about epistemic foundations and limitations of statistics and science. The event brought together researchers from diverse fields to discuss one of science’s most pressing challenges: The replication crisis. While the crisis is often attributed to systemic issues like “publish or perish” incentives, the discussions highlighted an overlooked culprit: a lack of understanding and acknowledgment of the epistemic foundations of statistics. After the workshop, our MCML members Lisa Wimmer, Moritz Herrmann and Patrick Schenk wrote a blog post with their thoughts on the topic.

MCML Director Daniel Rückert and his collaborators have achieved a groundbreaking milestone in medical AI. Their innovative model uses machine learning to precisely determine the timing of strokes based on CT imaging. This advancement is critical for improving stroke diagnosis and ensuring patients receive timely and effective treatment. By addressing a long-standing challenge in neurology, this research exemplifies the transformative potential of AI in healthcare.

Throughout the year, MCML researchers have offered workshops and lectures for school classes and participated in public events to talk about machine learning and artificial intelligence.

Our PI Björn Ommer has been awarded the prestigious Eduard-Rhein-Foundation Technology Prize 2024 for his contributions to latent diffusion models and text-to-image generators. This award recognizes his globally acclaimed innovations, which have significantly advanced the field.

The award ceremony will take place on Saturday, 24 May 2025 at 3 pm at Flugwerft Schleißheim (Hangar Schleißheim) in Oberschleißheim and honours outstanding achievements in science, business and culture.

Additionally, on Friday, 23 May 2025 at 4 pm, Björn Ommer will present his research at a colloquium hosted by the Technical University of Munich at Theresianum, Room 0606, TUM Main Campus.

Congratulations to Björn on this remarkable recognition!

Join the MCML team as a Science Manager for International PhD Collaborations and help run our PhD program, including planning events and workshops for the next generation of AI researchers.

Application deadline: 06.01.2025

Find more details on the TUM Job Portal.

In episode 128 of #digdeep, Dr. Julia Urbahn, founder of the Swiss market research institute Intervista, joins as the guest. Intervista specializes in qualitative surveys, price analyses, and other insights for both B2C and B2B markets.

The discussion focuses on the continued importance of traditional market research in an era dominated by digital behavioral data from platforms like Google. Dr. Urbahn emphasizes that understanding the why behind human actions is far more valuable than simply knowing the what.

She also highlights how digitalization benefits market research, noting that people tend to be more honest with digital surveys than with human interviewers. Drawing from her experience running Switzerland's largest digital panel, she provides a unique perspective on uncovering the motivations behind consumer behavior.

Just listen in (podcast in German).

MCML Director Daniel Rückert receives the Gottfried Wilhelm Leibniz Prize 2025. He is being honored for his research on AI-assisted medical imaging. The most important German research prize is endowed with 2.5 million euros by the German Research Foundation (DFG).

Congrats from us, Daniel!

MCML Junior Member Jesse Grootjen is writing his doctoral thesis at the Chair of our PI Albrecht Schmidt, Human-Centered Ubiquitous Media at LMU Munich. The group conducts research at the crossroads of human computer interaction, media technology, and ubiquitous computing. The overall research question is: how can we enhance human abilities through digital technologies?

On December 11th, we visited Munich Re for another event in our series „MCML Pitchtalks with Industry“ - this time, our focus was on generative AI, and how it transforms the reinsurance business.

After a presentation from MunichRe on open problems in generative AI, we heard exciting pitchtalks from some of our junior members: Leonardo Galli talked about Optimization. Silvia Casola presented her work in natural language processing. Zijian Ma and Yunpu Ma presented their work on AI Agents. And Zhicong Xian showed advancements in process mining. Exchange of ideas was vivid and informative - we are looking forward to visiting more companies within our community.

The Munich Center for Machine Learning (MCML) is highlighted in the latest issue of Bildung+ Schule Digital Magazine for its role in the innovative KITrans project.

KITrans is an innovative educational project bringing accessible AI learning formats to classrooms. Together with Prof. Uta Hauck-Thum from LMU Munich and Junge Tüftler, the project introduces students and teachers to AI's potential, challenges, and responsible use through interactive workshops and co-creative formats. Participants learn to develop and adapt AI models, gaining practical insights into algorithms and data analysis.

In a recent interview, our director, Daniel Cremers, highlighted how AI innovations aim to simplify and enhance daily life. His work emphasizes developing reliable, user-friendly AI systems while addressing ethical challenges and societal impacts. Prof. Cremers’ research in mathematics and machine learning helps shape AI's future as a positive force.

Our PI Matthias Nießner has been awarded the prestigious ERC Consolidator Grant for his research in computer vision and machine learning. His work focuses on developing advanced methods for 3D reconstruction, visual computing, and AI-driven content creation, with applications ranging from virtual environments to realistic simulations.

Congrats from us!

Our PI Nassir Navab has been honored with the ISMAR Career Impact Award. This prestigious recognition celebrates his exceptional contributions to the field of mixed and augmented reality, acknowledging his research, innovations, and lasting impact on this interdisciplinary domain.

Congrats from us!

MCML Junior Member Kevin Höhlein is a PhD student at the Chair of Computer Graphics and Visualization (TUM) of our PI Rüdiger Westermann, researching applications of data science and machine learning techniques in the context of meteorological data analysis.

The MCML PhD-Matchmaking portal has now closed, and applications are currently under review. We sincerely thank everyone who applied. This year, we were excited to receive an outstanding total of 1190 applications!

Stay tuned for future updates and announcements.

MCML Member Dominik Bär is a researcher at the Institute of Artificial Intelligence (AI) in Management at LMU, working within the research group of Stefan Feuerriegel. He is pursuing his PhD with a focus on social media analytics.

At OFF Radio in Poland, an experiment tested artificial intelligence as a radio host, using digital avatars of famous figures for interviews, including a virtual interview with Wisława Szymborska. The project was terminated after a week due to strong emotional reactions and collected feedback.

Our PI, Sven Nyholm, professor of ethics, emphasized on ZDF heute the importance of context and purpose when creating digital copies, along with transparency in their use. He also highlighted the need for consent, as seen with the Szymborska Foundation's approval for the interview.

We visited the Allianz Technology Campus in Unterföhring on November 25th, to have our second event in our series „MCML Pitchtalks with Industry“. Connecting dots between collaboration ideas and joint research interests on Generative AI and other hot topics was really fun and revelatory. After a presentation by Allianz on their AI research and topics, Shanshan Bai presented her work on Gen.AI and Earth Observation, Leonor Veloso on Gender Bias and NLP, and Anna Steinberg on how to analyze Greenwashing effects with AI.

The Data Science Certificate Program is a graded extra-occupational training program at LMU Munich designed for working professionals, researchers, and executives who want to improve their skills in the field of data analytics, statistics, and ML.

The program is offered twice a year. Apply now and join the training between March and June 2025!

In a new article published by TUM, 15 scientists from the institution are recognized as being among the most cited worldwide.

Among them are our PIs, Fabian Theis and Xiaoxiang Zhu, highlighting their significant contributions to their respective fields.

In the new article published at Research in Bavaria, the Munich Center for Machine Learning is highlighted as a crucial part of Bavaria’s AI and machine learning research.

It highlights the work of MCML’s members, who are advancing innovative solutions through collaboration and interdisciplinary research.

The article also discusses how MCML’s community helps strengthen Munich’s reputation as a leading hub for AI.

A third of shoes in Germany are bought online, yet sizing issues often lead to high return rates and waste.

Footprint, co-founded by Matthias Brendel, tackles this by allowing customers to measure their feet with a smartphone and matching them to the best-fitting shoes or models. This innovative solution reduces returns, cuts costs for retailers, and helps the environment. Matthias, who previously worked on innovations at Audi, shares his journey of founding Footprint and offers insights into building a successful startup.

Just listen in (podcast in German).

The Capital Magazine selected our PI Zeynep Akata as one of the Top 40 Under 40. In this interview, she shares insights insights into her groundbreaking work in explainable and robust artificial intelligence. She discusses how her research bridges the gap between complex AI systems and practical applications, particularly in healthcare, emphasizing the need for transparency and reliability in AI decision-making.

MCML General Assembly 2024 celebrated innovation and collaboration with inspiring talks, Junior Flash Talks, and meaningful networking opportunities.

The BMBF is funding two new joint projects with several MCML researchers involved. One teaches AI models causal relationships, the other refines the tactile abilities of robots.

CausalNet: AI that Understands Cause and Effect
The CausalNet project aims to advance machine learning (ML) by integrating causal reasoning, moving beyond current models that rely solely on correlations. This approach promises greater reliability and performance, particularly in fields like medicine, where understanding cause-and-effect relationships could enable targeted therapies.

Funded with nearly €2 million by the German Federal Ministry of Education and Research (BMBF), CausalNet will develop novel methods for embedding causality into ML, working with experts from LMU, TUM, KIT, Helmholtz AI, and Economic AI GmbH – including MCML PIs Stefan Feuerriegel (spokesperson of the project), Stefan Bauer, and Niki Kilbertus. The project will tackle challenges in high-dimensional environments using tools from representation learning, statistical efficiency, and specialized ML paradigms, with a focus on open-source outputs.

GeniusRobot: Enhancing Robotic Vision and Grasping with AI
The other project focuses on improving robotic manipulation using generative AI. MCML PIs Gitta Kutyniok and Björn Ommer are developing interpretable AI models that predict tactile information from visual data, enabling robots to dynamically adapt their grip.

The project leverages multimodal AI to integrate and interpret sensory inputs, enhancing robotic flexibility and resilience. It also explores converting tactile data back into visualizations, aiding manipulation of partially visible objects. The research aims to unlock new use cases in automated manufacturing and human-machine interaction, prioritizing safety and interpretability in critical environments.

Congrats to everyone involved!

With the company se3labs, we started our new event series „MCML Pitchtalks with Industry“ on November 13th. We had an insightful exchange of ideas on research that’s happening at the MCML. se3labs shared open questions in the field of computer vision and 3D reconstruction of complex environments, that se3 is currently working on. Our members Michael A. Hedderich, Yunpu Ma and Maximilian Bernhard shared their research on connecting questions - from Natural Language Processing, AI Agents, to Computer Vision.

MCML Junior Member Anna Reithmeir is writing her doctoral thesis at the Chair of “Computational Imaging and AI in Medicine” of our PI Julia Schnabel. Her work addresses the complexity of aligning images from different times or modalities, enabling the comparison of anatomical changes such as tumor growth.

MCML Principal Investigator and Bavarian AI Chair Gitta Kutyniok has been featured on Plattform Lernende Systeme.

Her profile highlights her pioneering research in the mathematical foundations of AI, bridging advanced theories with real-world applications. The impact of her work is profound in shaping the direction of AI research.

Two years ago, #digdeep became the first German-language podcast to welcome ChatGPT as a guest. Now, ChatGPT is back on #digdeep, showcasing its brand-new conversational capabilities—and it’s impressive!

This episode explores the origins of these advancements and their implications for business and jobs. It discusses progress in language processing, the role of AI in data analysis, and the challenges posed by workplace automation. It also delves into the future of collaboration between humans and machines.

As a bonus, it ends on a fun note: What happens when ChatGPT meets Siri? Spoiler—it’s hilarious!

Just listen in (podcast in German).

At this year’s TUM Talk for Mid-Sized Businesses, MCML Director Daniel Cremers, a renowned expert in computer vision and artificial intelligence, provided valuable insights into how digital innovations can reshape mid-sized enterprises.

The talk covered the practical applications of AI technology across various industries, highlighting how digital advancements drive competitiveness and growth for businesses.

MCML PI Alena Buyx, renowned German bioethicist, has been appointed as a new curator of the Bertelsmann Foundation, where she will contribute to strategic direction and foster dialogue on key societal issues.

With extensive experience in ethics and research, Professor Alena Buyx joins the Bertelsmann Foundation as its new curator. Her appointment marks an important step in advancing discussions on ethical challenges and sustainable solutions for the future.

Congrats from us!

The herCareer fair, with nearly 7,000 attendees, is one of the largest career events for women. We had the pleasure of recording a live episode of #digdeep on stage, enjoying the opportunity for direct interaction with our listeners.

The guest in episode 125 is Alexandra Renner, who serves as the link between BMW Garage and the development department. She connects innovation needs, smart ideas, and entrepreneurial minds. With her background as a co-founder right after her studies and her current role at BMW, Alexandra bridges the often contrasting worlds of corporate and startup culture. She shares fascinating insights into the challenges and opportunities of bringing these two realms together.

Just listen in (podcast in German).

This week, the MCML visited the Leibniz Rechenzentrum in Garching. Following a brief introduction to LRZ and their services for researchers, we received a tour of the LRZ supercomputer, including the MCML GPU cluster.

Thank you to the LRZ team for an insightful afternoon!

More than 1.000 visitors passed by the Science Communication Lab in the Museum during the Long Night of Munich Musuems. Our MCML researchers Mariia Gladkova, Tarun Yenamandra, Mateo de Mayo, Mia Brandtner and Qadeer Khan were on site providing insights on their research and answered all the questions related to AI.

With the AMELIE exhibit, the MCML invited visitors to learn about the basics of machine learning in a playful and interactive way. On a big, illustrated canvas we showed how AI works and explained the difference between AI and Machine Learning. The Canvas was designed by Laura Ekes.

In a second demo, we showed how it is possible to simultaneously localize and map the environment using a camera and a connected IMU sensor. The method is called DM-VIO and can be used in self-driving cars, for example.

We also provided an insight into Stable Diffusion, a deep learning model that generates images from text descriptions. We explained how Stable Diffusion learns the connection between image and text through a latent diffusion model process and invited the visitors to try out Stable Diffusion for themselves.

For the 25th year, 'More museum for everyone!' brought together over 100 institutions, creating a night filled with art, culture, science, and technology.

This year's All Hands meeting took place in Dresden, bringing together professionals from the field to share insights and collaborate. Participants had the opportunity to present their projects, engage in panel discussions, and attend interactive workshops that covered pressing topics in artificial intelligence. The event also offered networking opportunities, fostering connections and partnerships that encourage collaboration and innovation within the field.

Our Director Daniel Rückert has been honored with the prestigious MICCAI Society’s 2024 Enduring Impact Award, which recognizes individuals who have made significant and lasting contributions to the fields of medical image computing and computer-assisted interventions. This award celebrates his exceptional achievements and influence in advancing medical imaging technologies, shaping the field, and fostering innovation over the years.

Congrats from us!

Checkout our PIs Björn Ommer and Björn Schuller in the Terra X talk on AI: Artificial intelligence that recognizes our most secret desires, that protects us - and guards us. What new possibilities are opening up? And what dangers?

We’re thrilled and proud that two of our MCML PIs were selected by Harald Lesch and ZDF to share their expertise in this exciting feature!

Our PI Alexander Fraser is spearheading an important EU-funded project aimed at enhancing Large Language Models (LLMs).

His research project focuses on improving the integration of lesser-spoken languages into LLMs, the AI-driven technology behind ChatGPT and similar chatbot platforms. This initiative will not only advance the technology but also explore its societal implications, ensuring that these advancements benefit the community.

Congrats to Alex!

Checkout our PIs Björn Ommer and Björn Schuller in the Terra X short "Nobel Prize in Physics 2024: Harald Lesch reacts! (Björn too... and Björn too)"

We’re thrilled and proud that two of our MCML PIs were selected by Harald Lesch and ZDF to share their expertise in this exciting short feature!

Björn Schuller, our Principal Investigator at MCML, shares his insights on Terra X Harald Lesch, discussing how artificial intelligence can recognize our deepest desires and monitor our actions. This episode delves into the opportunities and potential dangers of AI in relation to human emotions.

The group of our PI Nassir Navab achieved remarkable recognition at MICCAI 2024, one of the world’s leading conferences in medical image computing and computer‑assisted interventions. Their outstanding research was honored with multiple prestigious awards across various tracks.

Congrats to the team!

Our PI Alena Buyx, an expert in ethics and public health, was awarded the prestigious Order of Merit during the 2024 German Unity Day celebrations. The award, presented by Federal President Frank-Walter Steinmeier, recognizes her significant impact on societal issues. This honor highlights Buyx's dedication to improving public health and advancing ethical standards.

Congrats from us!

The European Computer Vision Association (ECVA) has honored our PI Angela Dai with the Young Researcher Award 2024, recognizing her outstanding contributions to the field of computer vision:

Angela stands out as a pioneering researcher in 3D scene reconstruction and semantic scene understanding. Angela’s research has played a pivotal role in establishing modern 3D deep learning as a prominent and influential area of study. Her efforts have significantly broadened the scope of visual perception, transforming the landscape of research in the digitization and understanding of 3D environments.

Congrats from us!

Our PI Björn Ommer has won 2024 German AI Prize awarded by the newspaper WELT. This award in the category “Innovation” is given for groundbreaking achievements in the research and development of artificial intelligence.

The jury singled out the image generation system developed by Björn Ommer and his team for praise. Called Stable Diffusion, this system has democratized generative AI by becoming so compact and powerful as to remove the need for expensive large-capacity computers and being accessible to all through its open-source software.

Congrats from us!

For two months, the DSSGxMunich-Team supported two projects with the UN International Organization for Migration's Global Data Institute, and the National park administration Bayerischer Wald. Last week, they presented their results at the Bavarian Academy of Sciences and Humanities.

DSSGx Munich is organized by MCML PI Frauke Kreuter and MCML Director Bernd Bischl and their teams and funded by the Munich Center for Machine Learning.

A team of computational biologists from Helmholtz Munich has been honored with the prestigious Frontiers of Science Award at the International Congress of Basic Science (ICBS) in Beijing.

MCML PI Fabian Theis, together with Alexander Wolf, and Philipp Angerer received this recognition for their influential 2018 paper, “Scanpy: Large‑Scale Single‑Cell Gene Expression Data Analysis.”

The ICBS is a globally renowned event that gathers leading scientists, researchers, and scholars from across disciplines to showcase their latest innovations. The Frontiers of Science Award celebrates researchers whose work has significantly advanced their field — a distinction well earned by this outstanding team.

Congratulations from us!

In this episode, we take a ride in a Waymo robotaxi through the busy streets of San Francisco.

Since 2022, autonomous vehicles without safety drivers have been allowed to operate and transport passengers in the city. Waymo has about 300 of these cars on the road, packed with sensors and cameras, though the interior looks familiar—except for the empty driver's seat.

The ride was smooth and efficient, though surprisingly ordinary. While robotaxis are already in use, fully autonomous private cars are still some way off due to unanswered questions about business models and operational areas.

Just listen in (podcast in German).

Our Associate Johannes Maly was awarded the 2024 ACHA Charles Chui Young Researcher Best Paper Award. This annual award recognizes outstanding contributions by young researchers in the field of harmonic analysis, as selected by the editors and publisher of the Journal of Applied and Computational Harmonic Analysis (ACHA).

Johannes received this honor for his paper “Robust Sensing of Low-Rank Matrices with Non-Orthogonal Sparse Decomposition,” published in November 2023 (Vol. 67, ACHA).

Congrats from us!

MCML JRG Leader Benjamin Lange will be a speaker at this year's Digital Ethics Summit organized by the Rheinische Post (RP). He will join the panel "AI – Opportunities and Risks" as an expert.

MCML member Ludwig Bothmann, PostDoc in the group of Director Bernd Bischl is featured in a documentary by Bayerischer Rundfunk titled: "Zu viel los im Wald? Mit KI auf den Spuren von Tieren und Menschen".

In Bavaria, where there is a legal right to freely access nature, encounters between humans and animals frequently occur in forests. The documentary addresses the question of whether wildlife is suffering from the added stress caused by humans and whether animals still have sufficient quiet zones. To answer this question, AI will be employed to analyze footage from various wildlife cameras.

Ludwig Bothmann, as an AI expert, is helping to develop the model and explains some of the challenges the researchers face as well as some advantages of using an AI for this analysis.

The development of the Stable Diffusion generative AI technology has garnered our PI Björn Ommer a nomination for the prestigious Deutscher Zukunftspreis awarded by the President of Germany.

On November 27, 2024, it will be announced whether his project, "Democratizing Generative AI – Stable Diffusion from Development to Practice," will prevail against two other nominees. In this brief LMU Newsroom article, Björn Ommer also provides a concise explanation of what Stable Diffusion is all about.

Congrats from us!

Since 2011, the DAGM has annually granted the German Pattern Recognition Award for outstanding, internationally visible research in the fields of pattern recognition, computer vision, and machine learning. The German Pattern Recognition Award continues the tradition of the Olympus Award for Pattern Recognition (1992-2010).

In 2024, the German Pattern Recognition Award was granted to our PI Angela Dai for her outstanding scientific contributions in the area of 3D Scene Reconstruction and Semantic Scene Understanding.

Congrats from us!

Our PI Angela Schoellig was a guest on the podcast Deep Dive by MIT Technology Review to talk about the situation of robotics in Germany.

After presenting the new Robotics Institute Germany to the public this June, she is now talking about robotics in Germany and the impact of AI on robotics, among other things.

At ACL 2024, the premier conference in computational linguistics, the Senior Area Chairs of each track selected exceptional papers for the Area Chair Awards.

We are proud to announce that our PI Barbara Plank, together with Junior Members Leon Weber‑Genzel and Siyao Peng, and collaborator Marie-Catherine de Marneffe, received an ACL 2024 Area Chair Award for their paper: “VariErr NLI: Separating Annotation Error from Human Label Variation.”

Congratulations from us!

In this interview, MCML PI Reinhard Heckel discusses the importance of data in the context of AI. He explains why data is the crucial component for generative AI, which methods can be used to prevent biases, and how data protection should be approached in the context of AI.

Lern-Fair began as a response to the Covid-19 pandemic, connecting students with volunteer tutors to support their learning from home. It has since grown into a prominent organization with thousands of volunteers, aiming to ensure equal educational opportunities across Germany.

Elke Büdenbender, Germany’s "First Lady," has recently become its patron. Our junior member Maria Matveev, a founding member of Learn Fair, discusses the organization’s development and the role of digital tools and AI in enhancing education.

Just listen in (podcast in German).

DSSG 2024 – short for Data Science for Social Good – has officially started. Talented data scientists from all around the world arrived in Munich for the 2-month fellowship program that contribute to social initiatives in Munich in the summer months.

The kick-off event of DSSGxMunich 2024 took place in early August. During the event, both teams presented their projects to the audience, which they will be implementing together with their partners from nonprofit organizations:

• Together with the International Organization of Migration:
  Project 1: Estimating the Impact of Events on Migration Flows with Machine Learning - Enhance Foresight on Migration Drivers and Movements to Improve Data-Driven Allocation of IOM resources and to Provide Better Services to Migrants and Internally Displaced People.

• In collaboration with the Bavarian Forest National Park:
  Project 2: Predicting Visitor Flows in the Bavarian Forest - Support Sustainable Tourism and the Harmonization of Nature Protection and Travel in the Bavarian Forest National Park.

DSSGx Munich is organized by MCML PI Frauke Kreuter and MCML Director Bernd Bischl and their teams and funded by the Munich Center for Machine Learning.

The large-scale study conducted by our PI Stefan Feuerriegel and his team on political advertising on social media during the 2021 German federal election is the subject of this audio feature by Norddeutscher Rundfunk (NDR). MCML Junior Member Dominik Bär discusses the study's findings and their interpretation.

This short article focuses on the current work of MCML Director Daniel Rückert and his team, who studied the suitability of AI chatbots in hospitals.

While seeing potential in certain areas, the team concluded that chatbots would currently put patients' lives at risk. To minimize this risk, the scientists developed a method to test the reliability of future chatbots.

The TUM student councils honor outstanding lecturers in the subject areas of Electrical and Computer Engineering, Informatics, and Mathematics from the past two semesters. The special feature of these prizes is that they are awarded by the students themselves.

Our PI Matthias Althoff and his team were delighted to receive a TeachInfAward in the category "Best elective course in the winter semester 2023/24". They were honored for the lecture “Formal Methods for Cyber-Physical Systems”.

Congrats to the team!

MCML Junior Member Pedro Tomaz da Silva, PhD student in the group of our PI Julien Gagneur, and his co-authors have received the ISBM 2024 Best RegSys Poster Award for their poster "DNA language models reveal the architecture of nucleotide dependencies in genomes".

Congrats to Pedro!

We proudly announce that the Munich Center for Machine Learning and the appliedAI Initiative are joining forces to make the use of artificial intelligence in companies more efficient. Through the partnership, companies benefit from the latest AI research and practical application knowledge.

The first Munich AI Day on July 4, 2024, organized by the MCML, was dedicated to the latest developments and research in the field of Artificial Intelligence (AI) in the Munich Werksviertel district. Along with researchers, the AI Day brought together decision-makers from politics, business and the media.

Platform giants such as Amazon, Alibaba, Apple, and Google have become crucial gatekeepers in the economy, setting the rules for new business models. Many companies are trying to enter the platform economy but struggle to achieve relevance.

Prof. Felix Wortmann of the University of St. Gallen explores these challenges in his book "Der Plattform-Navigator" and discusses them on episode 122 of #digdeep. He highlights the diverse business models available and emphasizes the importance of the right mindset for success.

Just listen in (podcast in German).

A new BBC article addresses the question of whether AI is a potentially effective way to improve government services. To answer this question, the critical views of our PI Sven Nyholm on the accountability and reliability of AI are considered.

The International Academy of Medical and Biological Engineering (IAMBE) is composed of fellows who are elected in recognition of their exceptional and sustained contributions to the advancement of medical and biological engineering.

We are honored to announce that our PI Nassir Navab has been elected as a fellow of the IAMBE class of 2024.

Congrats from us!

Dr. Irène Kilubi, originally from the Democratic Republic of Congo, has built a remarkable career as a business engineer and consultant, working with top companies like BMW and Siemens.

Despite her challenging beginnings, she has become an award-winning speaker, influencer, and entrepreneur. Her focus is on societal and economic issues, such as Joint Generations and Community Building. Additionally, she serves as an Expert Advisor for the European Innovation Council Accelerator and is a university lecturer and sought-after keynote speaker.

Just listen in (podcast in German).

Our PI Stefan Feuerriegel and his team, in collaboration with the Politecnico di Milano and CENTAI from Turin, conducted a large-scale study on the topic of election advertising. In the run-up to the federal election, more than 80,000 political advertisements on Instagram and Facebook were analyzed. The algorithms seem to favor the contents of certain parties.

Founded in 2000 as the first academy for outstanding young academics, the Junge Akademie brings together exceptional minds from all scientific disciplines and the arts.

At this year’s annual ceremony, ten young scientists were welcomed as new members. We are delighted that our PI Niki Kilbertus is among them.

Congrats from us!

Our partners from the AI Competence Center LAMARR invite submissions of papers for this year's Symposium on "Scaling AI Assessments – Tools, Ecosystems, and Business Models".

Against the backdrop of the importance of Trustworthy AI, as well as its implementation and establishment, the event is primarily aimed at practitioners from the TIC sector, tech start-ups offering solutions to the above-mentioned challenges, and researchers from the field of Trustworthy AI. As the symposium will also discuss the EU AI Act and its legal dimensions, legal experts are also cordially invited.

Learn more about the latest research results of our Director Daniel Cremers.

He and his team have succeeded in creating precise 3D reconstructions using only images from two cameras. Previously, such reconstructions required many hundreds of perspectives. With this new method, it is possible to make significant advances in the preservation of historical moments as well as in autonomous driving.

Our PI Niki Kilbertus was awarded the Leopoldina Prize for Young Scientists for his research and teaching achievements.

As a Professor of Ethical Machine Learning, he focuses on the research and development of fair and transparent machine learning. As a pioneer in this field, he formalizes causal concepts of fairness so that they can be integrated into machine learning systems.

Congrats from us!

As part of this year's re:publica 2024 in Berlin, our PI Alena Buyx was part of the panel discussion "Data Protection vs. Data Usage: Irreconcilable Opposites?"

Together with moderator Theresa Sickert and the Federal Commissioner for Data Protection and Freedom of Information, Ulrich Kelber, she discussed the possible compatibility of data protection and data usage.

Follow the link to YouTube, if you want to learn more:

Our PI Björn Ommer was a guest on Gert Scobel's show to talk about AI in medicine and healthcare.

Together with the other guests, he discussed, among other things, whether the data on which AI systems learn is sufficiently diverse. Other topics included emerging ethical and data protection issues as well as the ways in which AI can and will be used in medicine.

This article highlights the work of our PI Nassir Navab on robotic ultrasound systems. By integrating such systems into everyday work, doctors are to be meaningfully relieved. However, to fully utilize the various advantages, Navab emphasizes raising patient awareness regarding the robots.

Our PI Sven Nyholm has a brief appearance in this newspaper article by Die WELT about the role of AI in India's political campaigns. He emphasizes the importance of legal restrictions on AI to protect and promote democracies.

Jazz is constantly evolving and defies easy definition, characterized by improvisation and interaction.

Arndt Weidler, deputy director of the Jazz Institute Darmstadt, highlights the institute's extensive jazz collection and its role in bridging science and practice. The digital revolution, particularly AI, is now impacting the jazz world, suggesting new possibilities for creativity and interaction. This raises the question of whether AI could become the new band member in jazz.

Just listen in (podcast in German).

For the 75th anniversary of the German Constitution, researchers from TUM and ETH Zurich stored the Basic Law as a DNA sequence and mixed it with ink. Our PI Reinhard Heckel was also part of this team and talks about the project and DNA as a storage medium in this short feature by Südwestrundfunk (SWR).

On April 18-19, the 18th International For..Net Symposium 2024 "Truth - Power – Rule of Law: Generative AI in the Mirror of Law, Technology and Society" took place.

MCML Junior Member Maria Matveev, a PhD student in the research group of PI Gitta Kutyniok, received the 2024 Engagement Award from the German National Academic Foundation (Studienstiftung des deutschen Volkes) in recognition of her outstanding social commitment as co-founder of www.lern-fair.de, a platform that promotes equitable access to education through volunteer-based tutoring.

Congrats from us!

In the context of the 105th German Radiology Congress, our PI Julia Schnabel was awarded the Alfred-Breit-Prize of the German Radiological Society (Deutsche Röntgengesellschaft) for her work in the field of radiological imaging.

Congrats from us!

We proudly announce that our PIs Stefanie Jegelka, Foundations of Deep Neural Networks and Suvrit Sra, Resource Aware Machine Learning have been honored with the prestigious Humboldt Professorship.

This award not only recognizes the achievements of our PIs but also advances the expansion of their research.

Congrats from us!

Generative KI, once hyped, now proves its transformative impact on various sectors, though still in nascent stages in many areas.

Vlad Varichev, from Accenture, navigates the integration of AI into corporate processes, noting the rapid evolution from sci-fi to reality, particularly in robotics. This journey into AI highlights its dynamic nature, promising continual evolution beyond current horizons.

Just listen in (podcast in German).

LMU’s research group in AI ethics, together with the MCML, the Center for Ethics and Philosophy in Practice, and the Oxford Uehiro Centre for Practical Ethics, welcomes submissions of extended abstracts (750 to 1000 words) for a two-day workshop on the ethics of partiality, human-robot relationships, and AI taking place in Munich this fall.

Rosmarie Steininger's experiences in the automotive industry taught her the importance of human connections, whether in BMW IT or in philanthropy, highlighting the effectiveness of bringing the right people together.

Leveraging the abundance of data available, she founded Chemistree, a startup aimed at connecting individuals such as mentors and mentees, recruiters and applicants, and scientists, driven by her passion to unearth buried talents and opportunities.

Steininger's commitment extends to initiatives like KIDD (AI in Service of Diversity) and the AI Standardization Roadmap, reflecting her dedication to leveraging technology for societal benefit.

Just listen in (podcast in German).

This LMU Newsroom article focuses on the research of our PI Stefan Feuerriegel in the field of AI in diagnosis and therapy. His team is working on a so-called Causal AI, which might be able to predict individual risk changes when taking medication.

The MCML AI Ethics research group invites submissions of abstracts (250-500 words) for a two-day conference on "The Ethics of Conversational Agents and Generative AI" taking place at LMU Munich this fall.

The conference will bring together academic researchers and industry AI Ethics practitioners interested in the ethical challenges raised by generative and conversational artificial intelligence.

Confirmed speakers include:

• Alena Buyx (MCML/TUM)

• Herman Cappelen (University of Hong Kong)

• Michael Klenk (Delft University of Technology)

• Jana Sedlakova (University of Zurich).

Abstracts on any topic relating to the ethics of generative AI and conversational agents are welcome!

Submission deadline: 15 May 2024, 1pm CET via mcmlaiethicsconference@lrz.uni-muenchen.de

The conference is organized by our PI Sven Nyholm and our JRG Leader Benjamin Lange.

Our Director Daniel Cremers was a guest at the AI Days in Paris, organized by the PRAIRIE Institute and France Digitale.

There, he presented, among other things, the MCML, its structure and activities. As part of the panel discussion, he spoke with Yann LeCun, Michael Jordan, Francis Bach, and Cordelia Schmid about the topic "Future of AI".

Marc Raschke has been lauded as one of Germany's most innovative communicators, earning praise from "Pressesprecher" and securing a spot on the shortlist of the 100 most influential figures in the PR and marketing industry by "W&V".

Despite the challenges posed by the pandemic in 2020, Raschke's adeptness in crisis communication was evident as he navigated the task of boosting morale within a financially challenged hospital, employing a blend of digitalization, social media engagement, and pragmatic strategies. Notably, his efforts during this tumultuous period garnered him a nomination for Communicator of the Year, underscoring his ability to adapt and thrive in crisis situations.

Additionally, Raschke's achievements extended beyond crisis management, as evidenced by his recognition as Research Spokesperson of the Year in 2021, further solidifying his reputation as a skilled communicator and problem-solver in the field of public relations.

Just listen in (podcast in German).

MCML Junior Member Lukas Gosch conducts research on how quickly - or slowly - AI algorithms react to changes. His field of work is the reliability of artificial intelligence, or more precisely, of machine learning algorithms. As these algorithms are playing an increasingly important role in our lives, he finds the sociological, political and ethical aspects exciting. In his research, he takes a look on what kinds of misbehavior of artificial intelligence exist.

MCML Junior Member Maximiliane Windl’s research focuses on understanding and mitigating privacy concerns in ubiquitous computing systems at the chair of our PI Albrecht Schmidt. Her main concern is one thing: how can people become more aware of what data these devices are constantly collecting?

Our Junior Research Group Leader, Benjamin Lange, was a guest on the Radio Stockdale Podcast hosted by Michael Sears. There, he discusses the fundamentals of Responsible AI governance, as well as its potential implementation. He advocates not only for compliance with certain rules but also for proactive and visible ethical leadership.

The MCML initiated a new collaborative format for scientific exchange and community building: Our Pitch Talk Series. Here, our junior members get the opportunity to pitch their research and connect within the MCML community.

At EACL 2024, the paper “Kardeş‑NLU: Transfer to Low‑Resource Languages with Big Brother’s Help – A Benchmark and Evaluation for Turkic Languages” received the Outstanding Paper Award.

The work, co‑authored by MCML PI Hinrich Schütze and Junior Member Lütfi Kerem Senel together with Benedikt Ebing, Konul Baghirova, and Goran Glavaš, introduces a comprehensive benchmark and evaluation framework to advance natural language understanding in low‑resource Turkic languages through cross‑lingual transfer.

Congratulations from us!

Our Junior Member Sarah Ball, a PhD student in the group of PI Frauke Kreuter, was a guest on the ARD AI podcast with Fritz Espenlaub and Gregor Schmalzried to discuss the credibility of AI. Sarah explains methods to make AI more trustworthy, ensuring that we as humans are not misled.

Ralf Niemeier, an entrepreneurial serial achiever, has garnered extensive experience in various sectors including family businesses and steel trading, preparing him to digitally revolutionize the logistics industry with his startup Visimatch.

Despite being based in Paderborn, not a typical startup hub, Visimatch focuses on developing platforms to address logistical complexities such as coordinating large-scale events and military exercises.

Niemeier emphasizes the importance of understanding and guiding customers through their digital transformation journey, highlighting the common mistake of digitalizers neglecting the practical needs of clients still entrenched in traditional methods like Excel.

Just listen in (podcast in German).

MCML Junior Member Lisa Wimmer researches the practicability of model predictions. By supplementing them with a reliable statement about their statistical certainty or uncertainty, she aims for improving the underlying models.

Artificial intelligence only becomes a disadvantage if it is overestimated or used incorrectly, says MCML Junior Member Martin Binder. In his research, he improves algorithms and, above all, simplifies the work of other scientists.

Will artificial intelligence replace the profession of journalism? Uli Köppen has a clear answer to the supposed bad news. She not only reports on AI and algorithms - she also uses AI and automation to digitize and develop journalism.

At Bayerischer Rundfunk, she heads the AI & Automation Lab and co-leads the BR Data team. For her, AI has become an indispensable tool for investigative research as well as for the development of journalistic products.

Just listen in (podcast in German).

The current issue of the TUM magazine "Faszination Forschung" takes a closer look at the work of our Director Daniel Rückert.

With his team, he strives to train medical AI in a way that ensures both patient data privacy and the impartiality of the AI in diagnosis. To achieve this, he employs "Differential Privacy", a method that adds a kind of statistical noise to the data.

MCML Junior Member Moritz Herrmann conducts research at LMU at the Institute for Medical Information Processing, Biometry and Epidemiology. He compares methods of AI and ML that could predict the survival time of cancer patients.

Besides his research, Moritz is very engaged as Transfer Coordinator at MCML. Here, he organizes and supports Open Science activities.

Our PI Stephan Günnemann was a guest on the science broadcast Forschung aktuell by Deutschlandfunk.

In the feature, he explains pruning and spartification, two techniques that can significantly reduce the computational power required for language models. This makes AI algorithms more effective and, above all, more economical.

Our Junior Research Group Leader Benjamin Lange was a guest on the podcast Digital Sensemaker hosted by Christoph Holz to discuss the tense relationship between profit and ethics.

He delves into topics such as the role of ethical consulting, its actual benefits in decision-making, and fundamental concepts linking computer science and philosophy.

In her research, MCML Junior Member Clara Sophie Vetter, PhD candidate in the group of our Director Daniel Rückert, wants to find out how depression and psychosis are linked - with the help of artificial intelligence. In the "Precision Psychiatry" working group at LMU Hospital, she is investigating how the brain changes when a person develops a mental illness. Her team is looking at the structure and function of the brain to find out whether it is possible to recognize illnesses earlier or predict their course. For her research, AI models were trained with data from patients who have a high risk of developing psychosis. The AI can then help to evaluate those large amounts of data and recognize patterns.

Clara is not only a dedicated PhD student at the MCML, she is also very involved within the MCML community as a Junior Representative. In addition, she regularly gives talks at schools to teach young people the basics and importance of AI.

MCML PI Volker Tresp organized an ELLIS Workshop together with Kristian Kersting (TU Darmstadt) & Paolo Frasconi (Università di Firenze).

The workshop concerned machine learning approaches which operate at the human abstraction level, where the world is described by entities, concepts, and their mutual relationships. Among other topics, multi-relational learning, learning with (temporal) knowledge graphs, and the extraction of statistical and logical regularities from data were covered. Methods included embedding approaches, graph neural networks, scene graph analysis, neuro-symbolic programming, and inductive logic programming.

The workshop brought together like-minded researchers who had fruitful interactions.

Episode 114 hosted Frank Riemensperger, a key figure in Germany's digital transformation and former chairman of Accenture Germany. He's known for his significant contributions to technology discussions, especially in digitalization and AI, through his work with acatech, BITKOM, and AmCham.

His insights into fostering innovation through collaboration and competition were particularly enlightening. Just listen in an inspiring conversation on the future of technology (podcast in German).

In this article from the Frankfurter Allgemeine Zeitung, our PI Stefan Feuerriegel discusses whether AI renders business administration redundant, and if not, how the field of business administration should evolve in the future.

Girls'Day 2024 is just around the corner! To support the professional growth of more women in technical professions, the MCML offers interested girls the opportunity to delve into the everyday life of AI researchers and data scientists.

To this end, the MCML, in collaboration with the TUM Think Tank and Women in AI & Robotics, is organizing a workshop on April 25, 2024. In this workshop, participants will work together with experts to acquire basic knowledge in the field of artificial intelligence. (Workshop in German)

In this interview, our PI Stephan Günnemann explains the important and central differences between the terms AI, automation and machine learning.

Furthermore, he discusses the fundamental technical aspects of artificial intelligence and shares his personal assessment regarding the development of AI technology in Germany.

Valentin Hofmann is actually a former MCML Junior Member. He was a researcher at the chair of our PI Hinrich Schütze and a PHD student at the University of Oxford and recently finished his PhD. He is now a Young Investigator (Postdoc) at the Allen Institute for AI. During his PhD, he also spent time as a research intern at DeepMind and as a visiting scholar at Stanford University. Valentin Hofmann’s research focuses on the link between natural language processing and the social and cognitive sciences.

Julian Rasch is an electrical engineer and computer science PhD at the chair of our PI Albrecht Schmidt. He is conducting research about the interaction of humans and machines in virtual space. His current work is about how to better communicate virtual movement intentions.

Our PI Nassir Navab was a guest at ZEISS Beyond Talks and discussed the enhancement of healthcare through Augmented Reality.

Professor Navab explains what the use of Augmented Reality in medicine looks like, to what extent it would alter everyday medical care, and how, in his opinion, the public should be introduced to this new form of treatment.

Researchers from LMU Munich have developed an AI model that can detect low blood sugar levels (hypoglycemia) in people with diabetes while they are driving.

The team around our PI Stefan Feuerriegel created a model, that only uses routinely collected driving data and head/gaze motion data. The model could be used to develop a real-time warning system that could help prevent accidents caused by hypoglycemia.

Since Artificial Intelligence has become increasingly prevalent in recent years, the LMU dedicates the newest episode of FokusLMU on how AI is impacting our professional, private, and socio-cultural lives.

In this lecture, our PI Björn Ommer discusses the applications of AI and the resulting social transformations. He addresses the opportunities and challenges posed by AI, and how it will shape our future.

Watch the presentation on YouTube.

Prof. Walter Radermacher, a leading German statistician, served as President of the Federal Statistical Office. He contributes to advancing statistics in various forums, including the UN and OECD, crucial for trust in data.

Statistics require explanation, representing approximations of facts. We need a reliable "label" to understand their validity and limitations. Yet, media often manipulates statistics to fit agendas, especially during major elections.

It's vital to ask: Why is statistics critical for a functioning democracy? Just listen in (podcast in German).

Our Director Daniel Cremers, Professor for Computer Vision and Artificial Intelligence at TU Munich, talks to Josef Griesbauer, Professor at Kempten University of Applied Sciences on his YouTube channel Der Griesbauer.

Together, they delve into discussions surrounding AI, machine learning, and the potential future implications influenced by AI. Their conversation covers a wide range of topics including the concept of intelligence, the philosophical notion of free will, and the significance of emotions in the context of AI.

The 2023 ACM Fellows were recently announced by the Association for Computing Machinery (ACM): MCML PI Albrecht Schmidt is one of them. Recognized for their pioneering contributions to computing science and technology, he joins a select class of ACM Fellows. Less than one percent of all worldwide ACM members are recognized as ACM Fellows for their outstanding achievements each year since 1993.

Congrats from us!

The 2023 ACM Fellows were recently announced by the Association for Computing Machinery (ACM): MCML PI Björn Schuller is one of them.

Recognized for his empirical and theoretical contributions to the development of computer audition, affective computing, and health informatics, he joins a select class of ACM Fellows. Less than one percent of all worldwide ACM members are recognized as ACM Fellows for their outstanding achievements each year since 1993.

Congrats from us!

In this guest article in the Neue Zürcher Zeitung (NZZ), our Junior Research Group Leader in AI and Ethics, Benjamin Lange writes about AI-generated deepfakes and the limited possibilities of detecting them with AI tools.

Instead of relying solely on technical aids, he also calls for resilience, education in the field of AI, and a critical attitude to sharpen our perception.

We proudly announce that our PI Björn Ommer has been appointed as a member of the Bavarian AI Council. The AI Council was established by the Bavarian state government in 2020 and has since served as an advisory body to the state government.

Professor Ommer caused an international sensation in 2022 with the development of the diffusion model that serves as the basis for Stable Diffusion. This makes him one of the top international scientists in the field of artificial intelligence.

His expertise makes a decisive contribution to Bavaria as a high-tech location.

Dr. Enise Lauterbach, ex-cardiologist, founded LEMOA in 2019 to detect atrial fibrillation through data, aiding heart failure patients. Her switch to entrepreneurship, unusual for doctors, garnered awards and recognition.

We explore her transition and colleagues' responses. Just listen in (podcast in German).

The Science Communication Lab at Deutsches Museum hosted our MCML exhibition that attracted thousands of visitors over three weeks.

The exhibition featured interactive exhibits on machine learning, simultaneous localization and mapping (SLAM), and Stable Diffusion, an algorithm that connects images and text. Visitors gained insights into cutting-edge research and enjoyed engaging hands-on experiences.

Our PI Alena Buyx was a guest on the SWR podcast "Die Schule brennt" to discuss the challenges facing the education system in dealing with AI.

"AI must not replace humans", Professor Buyx clarifies in a statement by the German Ethics Council. Against this backdrop, Professor Buyx and the moderator talk for almost 48 minutes about the advantages and disadvantages of AI in the education system. The use of AI offers opportunities, such as the promotion of individual learning or fair access to education, as well as risks.

Data Science for Social Good is a 2-months summer program in which 2 teams of fellows work on real-world data science projects to make a meaningful impact for the social good.

DSSGx Munich is organized by MCML PI Frauke Kreuter and MCML Director Bernd Bischl and their teams and funded by the Munich Center for Machine Learning. Apply now!

In this article our PI Hinrich Schütze shares insights into the challenges of automatic translation using AI technology.

The focus lies on the difficulty of training AI models to handle various languages, especially those with limited training data. Professor Schütze also emphasizes current issues such as bias and hallucinations in AI, working on integrating a "Working Memory" to enhance factual accuracy in responses and enable more effective use of AI in language processing.

«In our field, we’re living in fascinating times. Suddenly, intelligences have developed that nobody can really explain.»
(H. Schütze)

As CTO of Trumpf, a German engineering leader, Tom Schneider addresses industry evolution by emphasizing adaptation to sensor tech, data analytics, and new business models. Success hinges on embracing transformation and taking calculated risks, insights he shares for firms like Trumpf in navigating future challenges (podcast in German).

In the 10th episode, our MCML PI Sven Nyholm and his co-host John Danaher answer questions from podcast listeners. Topics covered include: the relationships between animal ethics and AI ethics; religion and philosophy of tech; the analytic-continental divide; the debate about short vs long-term risks; and much much more.

As we celebrate the season, may your algorithms be efficient, your datasets be abundant, and your Christmas be full of exciting data points.

Happy holidays from the Munich Center for Machine Learning!

We look forward to seeing you in 2024!

Professor Monika Schnitzer chairs the "Council of Economic Experts", Germany's leading economic advisory body. In the 2023 recap podcast she discusses the nation's digitalization lag and the imperative for educational reform. Just listen in (podcast in German).

Explore our MCML-Exhibition at Deutsches Museum Munich. The exhibition offers an insight into current research on Artificial Intelligence, with a focus on the field of Computer Vision.

We look forward to seeing you there!

This SZ article presents the position of our PI Björn Ommer on the ongoing scaling of AI models via computing power. Contrary to the general trend, Ommer's approach is to make models smaller but smarter. This is not only about practicability, but also about saving valuable resources.

As part of the Deutsches Museum's "Wissenschaft für Jedermann" series, our Director Daniel Cremers, our researcher Julia Moosbauer and our PI Matthias Nießner gave talks on the topic of AI.

The EU strives for AI regulation, emphasizing ethics amid rapid development. AlgorithmWatch, founded in 2016, evaluates the EU's AI Act.

In episode 109, Matthias Spielkamp discusses achievements, challenges, and the complexity of shaping AI policy amidst diverse interests. Just listen in (podcast in German).

In the 9th episode, our MCML PI Sven Nyholm and his co-host John Danaher explore the future of humanity's tech relationship. Will increased integration lead to greater dependence? They ponder the normative and axiological implications, delving into the ethical aspects of technology and the value of speculative discussions on the future.

The current white paper from the PLS working group led by our PI Volker Tresp uses practical examples to illustrate the opportunities and challenges of language models and analyzes the conditions under which companies can leverage the potential of the technology with confidence and legal certainty.

The experts recommend the creation of an open, commercially usable data set in German that complies with European values and rules and supports the development of language models in Germany.

«From the perspective of AI research, LLMs represent a significant technological breakthrough. They unlock the intelligence of language. They enable solutions that were previously beyond the realm of technological possibilities, but now seem to be taken for granted.»
(V. Tresp)

In episode 108, guests Max Haberstroh and Christian Kohlschein discuss transitioning from research to Accenture, emphasizing the impact of models like GPT-4 on data analytics in industrial development. The conversation underscores challenges in data quality and management for effective industrial application. Just listen in (podcast in German).

Our PI Alena Buyx was guest on the television show Studio 3 - live aus Babelsberg on RBB.

In this conversation, she discusses her most challenging moments — such as the debate on medical triage — and reflects on how ethical reasoning can guide us through today’s pressing issues, including the growing influence of artificial intelligence on our collective future.

In this interview our PI Stefan Feuerriegel talks about the danger of AI generated fake news. He explains what makes these campaigns so dangerous and how to counter them.

In the 8th episode of their podcast, our MCML PI Sven Nyholm and his co-host John Danaher talk about relationships with machines. Can you collaborate with a machine? Can robots be friends, colleagues or, perhaps, even lovers? These are just some of the questions addressed in this episode.

Episode 107 features Christof Kellerwessel, former Ford Electronics Chief, discussing the automotive industry's shift to software-driven products. The conversation goes beyond cars, exploring the human response to change - both in the industry and personal leadership. Just listen in (podcast in German).

Every year the US company Clarivate publishes a list with the most cited scientists in their specific research areas or overall in different scientific areas.

Our PIs Fabian Theis and Xiaoxiang Zhu are featured in their respective research areas.

Congratulations from us!

In this interview our PI Björn Ommer shares his opinion about the current development of Large Language Models. He talks about the problematic of computing power in AI, the importance of data security and how Germany could potentially solve these problems.

In this workshop organized by BR (bavarian public broadcasting) researchers show each other the projects and methods they are working on in short pitches and subsequently explore possible collaborations on use cases and methods.

Our PI Barbara Plank explains how it is possible to differentiate between machine-generated and human-written texts despite ever-improving AI. She also explains what is now important in the further development of language models.

Check out the new LMU Newsroom article about how LMU is utilising the possibilities of AI in various areas and how AI is changing teaching and learning. The article features our PIs Frauke Kreuter, Gitta Kutyniok, and Björn Ommer, as well as some of our PhD candidates. The latter talk about their own experience with AI at LMU.

Hosted by the AI Section of the Gesellschaft für Informatik (GI) the program has been around for about 20 years. This year the event offers an exciting lineup of courses, workshops, and activities that span various disciplines and promote collaborative, innovative thinking. The focus topic 2024 is Resilience, Robustness, Responsibility.

This winter semester (WS23/24), LMU launched a minor in Artificial Intelligence with the aim of offering a tailor-made education in artificial intelligence for non-computer scientists.

Starting in the winter semester 2023/24, LMU is introducing Artificial Intelligence as a major Minor — an AI program tailored specifically for undergraduates studying subjects other than informatics.

This new subsidiary subject is part of the AIM@LMU program within Germany’s National-Regional Development Initiative “Artificial Intelligence in Higher Education.”

The program was made possible thanks to the proposal by our Director Bernd Bischl and PI Eyke Hüllermeier.

In the 7th episode of their podcast, our MCML PI Sven Nyholm and his co-host John Danaher talk about the moral status of machines and discuss wether machines can be more than just mere things. Sven and John share their thoughts on this topic and give some guidance as to how to think about the nature of moral status and its significance.

The streetscape is changing. Cars with electric number plates are commonplace, charging is the new refuelling, and on the motorway you can see drivers taking their hands off the steering wheel. Marvin Ruf works for the company HERE. He shares his opinion about the current and future development of mobility - just listen in (podcast in German).

Join the team of highly motivated and aspiring researchers, led by our PI Stefan Feuerriegel.

In the 6th episode of their podcast, our MCML PI Sven Nyholm and his co-host John Danaher discuss the controversy arising from the idea moral agency in machines. What is an agent? What is a moral agent? Is it possible to create a machine with a sense of moral agency? These are just some of the questions up for debate.

Our PI Julien Gagneur, together with researchers from Sweden and Belgium, received a Synergy Grant from the European Research Council (ERC) for their research on deciphering the language in which our genomes are written. In this interview, they talk about their project EPIC and what the synergy grant means for them and their future work.

Congrats from us!

Our PI Stefan Feuerriegel argues why more AI should be build around open-source principles to be more accessible to everyone.

Law & AI is the stuff of many science fiction movies: AI solves crimes, or prevents them before they happen. But the reality relies on analog legal systems grounded in human processes. How will the interaction between AI and the legal system look like? Tom Brägelmann shares his opinion in episode 105 - just listen in (podcast in German).

In the 5th episode, our PI Sven Nyholm and his co-host John Danaher delve into responsibility gaps in technology. Over two decades, legal and philosophical research has addressed the challenge posed by increasingly autonomous machines. They examine the significance, ethical and legal implications, and potential resolutions for this issue.

Aligned with the non-profit Carpentries' mission, the course enhances didactic and teaching skills. Participants can broaden their ability to share coding and data knowledge, fostering accessibility to these skills for the general public.

In episode 104, Professor Enzo Weber challenges the view of Germany lacking digital infrastructure. As a macroeconomist at IAB, he provides nuanced insights, evaluating the country's digital standing since the millennium. Despite media skepticism, he discusses Germany's strengths in the digital economy. Tune in for concise analysis (podcast in German).

In the 4th episode, our MCML PI Sven Nyholm and his co-host John Danaher talk about the role that technology can play in changing our behaviour. In doing so, they note the long and troubled history of philosophy and self-help. They also ponder whether we can use technology to control our lives or whether technology controls us.

On October 9 and 10, the All Hands Meeting 2023 of the German AI competence centers (BIFOLD, ScaDS.AI, Lamarr Institute, Tübingen AI Center, DFKI and MCML) took place in Berlin.

As one of a total of five university competence centers, the MCML accepted the invitation to Berlin to present and exchange current research results on the EUREF campus and to discuss them afterwards.

The event was a success all along the line. It is important to see that Germany is doing AI research from top level and that the state wants to promote and advance this fact. We thank all participants and especially BIFOLD for organizing this great event.

The Data Science Certificate Program is a graded extra-occupational training program at LMU Munich designed for working professionals, researchers, and executives who want to improve their skills in the field of data analytics, statistics, and ML. The program is offered twice a year. Apply now and join the training between February and June 2024!

Our PI Björn Ommer is featured in a new post by Forbes Magazine. The article looks at why bigger is not always better in generative AI. Follow the link for the whole article.

In the 3rd episode, our MCML PI Sven Nyholm and his co-host John Danaher discuss risk and technology ethics. They focus, in particular, on the perennially popular and widely discussed problems of value alignment and control. They start the conversation with the famous case study of Stanislov Petrov and the prevention of nuclear war.

At the end of the two-months program, the closing event of the DSSGx Munich 2023 took place on 28 September. As the highlight of the event, our two teams, consisting of eight fellows from all over the world, each presented their projects.

DSSGx Munich is organized by MCML PI Frauke Kreuter and MCML Director Bernd Bischl and their teams and funded by the Munich Center for Machine Learning.

Our Data Science for Social Good (DSSG) fellows had the privilege of visiting the Bundeskanzleramt and meeting members of the German Parliament in Berlin.

Our team got a tour at the Bundeskanzleramt and a presentation of the newly founded Datalab. In addition, the fellows and the rest of the DSSGx Munich 2023 team had the honor to meet and present their projects to important personalities of our parliament.

DSSGx Munich is organized by MCML PI Frauke Kreuter and MCML Director Bernd Bischl and their teams and funded by the Munich Center for Machine Learning.

Our MCML PhDs Theresa Ullmann and Julian Lange received the "GMDS Förderpreis 2023" of the German Association for Medical Informatics, Biometry and Epidemiology (GMDS) for their outstanding theses.

Theresa received the award for her PhD thesis, Julian for his master thesis, both supervised by our PI Anne-Laure Boulesteix.

Congrats from us!

Our MCML PIs Bernd Bischl, Stephan Günnemann, and David Rügamer and MCML PhD Lisa Wimmer received the Best Paper Award for the Research Track at the ECML PKDD 2023 with their paper "Towards Efficient MCMC Sampling in Bayesian Neural Networks by Exploiting Symmetry". The paper is a joint work with Jonas Gregor Wiese and Theodore Papamarkou.

A new empirical study by the group of MCML PI Stefan Feuerriegel analyzes the workings of russian bots on X/Twitter in relation to the war in Ukraine. It was featured in a new article by the German magazine FAZ.

The MCML is organizing together with impACTup!, Netlight Consulting GmbH and ZD.B / LMU Media Informatics Group this two-part event, where we aim to shift researchers and students in ML and AI into an entrepreneurial mindset. This edition of our event aims to unlock the potential of start-ups utilizing AI to ensure food security.

Studio guest Niko Abramidis turns digital into art and art into digital.

So many fascinating questions: What does authenticity mean? Does the blockchain have a physical counterpart? Why is art a time machine? (podcast in German).

The german television network ZDF aired a programm on the opportunities and risks of AI in the medical sector, interviewing MCML PI Julia Schnabel as an expert in this field.

In this insightful piece, our PI Barbara Plank elucidates the intricacies of Natural Language Processing, an interdisciplinary research field central to AI that revolves around the computerized processing and generation of human language.

In this comprehensive and insightful article our PI Björn Ommer explains his research in the field of computer vision and the impact of generative AI on creativity and the arts. Furthermore, it provides insight into the decision to make Stable Diffusion open source.

The MCML and LRZ (Leibniz Computing Center) cooperate closely in the field of hardware, also due to the engagement of MCML Director Thomas Seidl in the board of the LRZ. In this article, Dieter Kranzlmüller, director of the Leibniz Computing Center, shares insights on cybersecurity in the era of the first quantum computers.

Studio guest Thomas Andrae is an investor and digitalization thought leader who also has a passion for art and culture. Tune in as we explore topics ranging from startups to preparing the youth for a digital future (podcast in German).

For her project entitled "Spatial 3D Semantic Understanding for Perception in the Wild" our PI Angela Dai has been awarded an ERC Starting Grant 2022 by the European Research Council (ERC).

Angela Dai's research focuses on attaining a 3D understanding of the world around us, capturing and constructing semantically-informed 3D models of real-world environments. This includes 3D reconstruction and semantic understanding from commodity RGB-D sensor data, leveraging generative 3D deep learning towards enabling understanding and interaction with 3D scenes for content creation and virtual or robotic agents.

Congrats from us!

For two months, two teams of 4-5 fellows from across the world will work on machine learning, artificial intelligence, and data science projects that have a social impact.

Both teams will be assisted by a Technical Mentor and a Project Manager.

DSSGx Munich is organized by MCML PI Frauke Kreuter and MCML Director Bernd Bischl and their teams and funded by the Munich Center for Machine Learning.

The MCML had the privilege of hosting a delegation led by PStS Mario Brandenburg from BMBF and Amtschef Dr. Rolf-Dieter Jungk from BayStMWK. Eminent MCML researchers presented the center's cutting-edge research across various domains. The event concluded with insightful discussions during the poster reception.

100 episodes of #digdeep and seven years online - after the men's round, we continue this time with our female guests. A third of our studio guests have been women so far, and we're taking a wild ride through the digitization garden in this episode as well (podcast in German).

Each year the Association for Computational Linguistics awards papers for their long-lasting impact on the field of Natural Language Processing and Computational Linguistics. This year the paper "Automatic Word Sense Discrimination" of MCML PI Hinrich Schütze, which was originally published in 1998 by MIT Press, got awarded.

Congrats from us!

Authors are invited to submit full-length high-quality papers as well as extended abstracts related to the topic of scene graphs and graph representation learning.

The Munich Center for Machine Learning invites you to a talk by Professor Jürgen Schmidhuber as part of the Munich AI Lectures. He is one of the best-known researchers in the field of artificial intelligence and will talk about his "Thoughts on Artificial Intelligence" on July 14, 2023 at 3 pm in the Audimax of Ludwig-Maximilians-Universität München (in German).

In our anniversary episode 100, we take a journey through the development of the last years: from #Bot to #ChatGPT, from #Facebook to #Algorithmwatch, from #Covid to #Agile (podcast in German).

On 30th of May, MCML cooperated with the Deutsche Jounalistenschule by organizing a workshop that brought together young journalists and AI junior researchers. As a follow-up to the event, our PI Stefan Feuerriegel, LMU professor for AI in Management at MCML, gave an interview about the impact of AI on media and journalism.

Our PI Björn Ommer, head of the Computer Vision & Learning Group, which developed Stable Diffusion, talked in an interview with the German Manager Magazine about OpenAI, the importance of computing power in the EU, Europe's place in current AI developments and its impact on jobs.

This year, AI for People is organizing the conference "AI for People: Democratizing AI (CAIP’23)". The conference will be an occasion to unpack and critically examine what "democratizing AI" really entails. The Call for Papers is now open, and you can send your papers in until the 14th of July.

The LMU Open Science Center is inviting masters and doctoral researchers to join a 5-day summer school so they can gain more trust in their research, and make it as credible as it can be in the eyes of their peers, the public, and funding agencies.

Our PI Frauke Kreuter is one of the main organizers of the WiDS Munich Conference. WiDS is an independent event to coincide with the annual Global WiDS Conference held at Stanford University and 150+ locations worldwide. All genders are invited to attend WiDS regional events, which features outstanding women doing outstanding work.

Our PI Björn Ommer, head of the Computer Vision & Learning Group, which developed Stable Diffusion, gave a talk at re:publica, Europe's biggest festival for digital society.

He talked about the future possibilities of generative AI, emphasizing on the need for a social dialogue, its dangers and how to maximize its profit.

Our PI Volker Tresp is one of the authors of the new White Paper of Platform Lernende Systeme (PLS). The experts from PLS believe that research in Germany is in a good starting position to tackle the technical challenges of AI technology, develop it in line with European values and thus exploit its potential for Germany and Europe.

The prerequisite: a European ecosystem for the development of large language models as well as application-oriented research that facilitates the transfer into industrial practice. The current white paper from the platform identifies the most urgent research needs and options for action.

«If we want to use language models for applications in and from Europe, we need European language models that master the local languages, take into account the needs of our companies and the ethical requirements of our society.»
(V. Tresp)

Technologies that improve human capabilities change risk behaviour, shows a new study led by our PI Albrecht Schmidt.

Daniel Rückert is Alexander von Humboldt Professor for AI in Medicine and Healthcare at the Technical University of Munich. He is also a Professor at Imperial College London. His field of research is the area of Artificial Intelligence and Machine Learning and their application to medicine and healthcare.

HHAI aims for AI systems that assist humans and vice versa, emphasizing the need for adaptive, collaborative, responsible, interactive and human-centered intelligent systems that leverage human strengths and compensate for human weaknesses, while taking into account social, ethical and legal considerations.

Our Junior Member Dominik Bär has been recognized for his AI research in the field of social media. Bär wants to use AI to solve social problems and improve social media.

«My research leverages methods from AI and machine learning to study social issues. Here, we use large-scale data from social media, for example, to detect conspiracy theorist, study the spread of misinformation, or analyze the impact of social media on elections.«
(D. Bär)

Congrats from the MCML!

It was great to meet the next generation of smart and enthusiastic young people who are interested in pursuing careers in the field of ML. It's important to encourage and support talents from a young age to pursue their interests in these fields, and to break down any existing barriers!

The MCML partnered up with Women in AI & Robotics to invite young and motivated talents to this year’s Girls' Day at LMU. 12 preselected female talents will have the opportunity to dive into the fascinating world of AI for a day and get an insight into the work of female scientists at MCML. We are excited to welcome the new female generation in AI!

Technological developments constantly present us as a society with new kinds of questions - and they are faster than social discourse or legal regulation. Our studio guest Professor Judith Simon is one of 26 members and tells us how the Ethics Council sees the relationship between humans and machines.

The annual Data Fest is promising to be an exciting event for data enthusiasts. The DataFest Germany is a 48-hour competition where teams of Bachelor and Master students from different disciplines come together to analyze a large and complex data set provided by an organization or company. The MCML is looking forward to seeing you there!

The MCML took part at ai.Bay 2023, the Bavarian International Conference on AI, bringing together experts from all over the world to exchange knowledge and ideas on the latest developments in AI. The two directors of MCML, Thomas Seidl and Daniel Cremers, along with other MCML PIs delivered exciting talks throughout the conference.

The German Data Science Days bring together data scientists from various fields creating a forum for users and experts. The MCML participated with a talk from our Director Bernd Bischl and a great session including our PI Fabian Theis on Data Science, ML and AI in university education.

State-of-the-art machine translation is trained on large parallel corpora, texts and their translations. But such corpora are not available for less-resourced languages.

With Data4ML our PI Alexander Fraser wants to provide a system for the rapid and inexpensive creation of new parallel corpora that can be used by the less-resourced language community themselves.

Congrats from us!

In the studio is science journalist Andreas Loos from Zeit Online. He has won many awards for his research and articles on data and digitization, including the Grimme Online Award. But is his work and expertise still needed at all, now that AI-based text generators and chatbots have taken over the opinion supremacy?

Read how our PI Stefan Feuerriegel uses AI to better evaluate global development aid projects.

DSSG is an initiative founded at the University of Chicago in 2013. Its mission is to create programs, solutions and communities that enhance the use of data science and AI for social good. Through the collaboration of the department of statistics at LMU Munich and the MCML we are now able to launch our own version of the DSSG fellowship program.

DSSGx Munich is organized by MCML PI Frauke Kreuter and MCML Director Bernd Bischl and their teams and funded by the Munich Center for Machine Learning.

BR24 quotes, among others, our PI Björn Ommer, who heads the "Computer Vision & Learning Group" at LMU Munich. This is where "Stable Diffusion" was developed, one of the most exciting AI applications in 2022, which transforms texts into images.

The MCML is organizing together with impACTup!, Netlight Consulting GmbH and ZD.B / LMU Media Informatics Group this two-part event, where we aim to shift researchers and students in ML and AI into an entrepreneurial mindset.

Join one of the top AI research centers in Germany and in Europe with access to an excellent network of scholars and scientists and lead a research group for which you are expected to hire two PhD students, who will be funded by MCML.

The GPT-3 language model is now available in a version optimized for conversations. ChatGPT from Open.ai is available to everyone in a trial version and shows where the journey is headed. We invited ChatGPT to our podcast and talk to the algorithm: So our studio guest is an AI bot (podcast in German).

Within the Thomas Bayes Fellowship Program of the Munich Center for Machine Learning, LMU Munich seeks to appoint a highly qualified junior academic to this professorship and, therefore, especially encourages early-career scholars to apply.

MCML PI Fabian Theis has won the 2023 Gottfried Wilhelm Leibniz Prize. The full professor for Biomathematics at TU Munich has been honored for his pioneering work in the analysis, modeling and interpretation of genomic data.

The Leibniz Prize, the most important research prize in Germany, is endowed with 2.5 million euros.

Congrats from us!

MCML PI Göran Kauermann and Giacomo De Nicola from the LMU Department of Statistics, together with Michael Höhle from the Robert Koch Institute, have been awarded the Corona Special Prize of the Federal Statistical Office for two papers on estimating excess mortality in the pandemic years 2020 and 2021.

Congrats from us!

The Munich Center for Machine Learning of the Munich universities is considered an important Bavarian competence center for artificial intelligence. To drive innovation, it is looking for companies to collaborate with.

The Munich Center for Machine Learning (MCML) in collaboration with the Munich Science Communication Lab (MSCL) is looking for a Science Communication Coordinator (m/f/d) starting as soon as possible.

Industrial AI Podcast spoke with our three directors Bernd Bischl, Daniel Cremers, and Thomas Seidl about research priorities, the battle for the best minds and the role of tech companies.

They also asked what academics should do in the field of AI in the future and how they deal with competition from Meta, Alphabet or AWS.

The Chair of Database Systems and Data Mining at the Institute of Informatics at LMU and the Munich Center for Machine Learning is looking for an IT System Administrator (m/f/d) starting February 01, 2023.

MCML PI Björn Ommer invited on October 6, 2022 artist and researcher of the Angewandte Wien Pamela Breda and students of the TUM Computer Vision & Artificial Intelligence Chair of our Director Daniel Cremers accompanied by EuroTech Marie Curie Fellow Roman Pflugfelder to his group.

The MCML is happy to promote this Hackathon, a collaboration of Women in AI & Robotics, the University of Waterloo Department of Mechanical and Mechatronics Engineering, and Mila.

You will work as a part of a team to develop your artistic robotic application on a three days journey between tutorials, coding, application development, and talks from robotics researchers.

This is a fun way to introduce robotics to curious minds.

MCML PhD Theresa Ullmann won the Best Poster Award at the 31st International Biometric Conference (IBC 2022) with her paper "Over-optimism in unsupervised microbiome analysis: Insights from network learning and clustering".

The paper is a joint research work of Theresa Ullmann, Stefanie Peschel, Philipp Finger, Christian L. Müller and Anne-Laure Boulesteix.

Congrats from us!

For five years, "Platform Lernende Systeme" has been promoting interdisciplinary exchange and social dialog on artificial intelligence in Germany. At the core of the platform are its seven working groups. MCML PI Volker Tresp is new co-head of the working group "Technologies and Data Science".

MCML PI Hinrich Schütze received a Proof of Concept Grant for his project Respond2Hate (Responsive classifiers against hate speech in low-resource settings).

In his project, Schütze is seeking to develop a pilot browser extension that allows users to locally remove hateful content from their social media feeds themselves – including in countries with little representation in current training datasets. To this end, he will validate self-developed adaptive models, continuously enhanced by means of natural language processing and deep learning techniques, for the filtering of hate speech.

Congrats from us!

The Machine Vision Learning Group led by our PI Björn Ommer has developed one of the most powerful image synthesis algorithms in existence.

Creating images from text in seconds – and doing so with a conventional graphics card and without supercomputers? As fanciful as it may sound, this is made possible by the new Stable Diffusion AI model. The underlying algorithm was developed by the Machine Vision & Learning Group led by our PI Björn Ommer.

When it comes to AI-based text-to-image generators, the spotlight has recently focused on Stable Diffusion, a project in which MCML PI Björn Ommer's research group is involved. It is probably the most powerful AI text-to-image generator to date, freely available to Internet users.

With Stable Diffusion, an AI system is freely available that generates impressive images. MCML PI Björn Ommer's research group is involved in the project.

The ERC has published a "Results in Brief" article on MCML PI Alexander Fraser's ERC Starting Grant DASMT: Our globalised interconnected reality demands ever-smarter automatic translation tools. Via deep learning, Fraser's team delivers solutions for statistical machine translation.

Congrats from us!

MCML became a permanent research institution in July 2022, funded by the German BMBF and the Bavarian HighTech Agenda.

We celebrated this milestone at our MCML Kick-Off Event with our members. The agenda included welcome addresses and selected short presentations from academia and industry. In addition, the event focused on networking with representatives from politics, industry, and academia.

Our two MCML Directors, Bernd Bischl and Daniel Cremers, recently gave an interview to Süddeutsche Zeitung about the future of artificial intelligence research in Munich and the growing role of the Munich Center for Machine Learning (MCML).

The Munich Center for Machine Learning (MCML), a joint initiative of LMU and TUM, has been granted funding as a permanent establishment following a positive evaluation.

«MCML offers a very attractive scientific environment with excellent opportunities for cooperation. Obtaining permanent funding from the federal and state governments is a great success and testifies to the outstanding quality of MCML. It enables participating scientists in Munich to further advance their machine learning research projects.»
(Professor Bernd Huber - LMU President )

«The decision by federal and state governments to make MCML a permanently funded center of excellence is a clear marker of the success of our One Munich strategy. By pooling our strengths, we want to drive forward future development in the domain of artificial intelligence. This will make Munich an even stronger magnet for young talents.»
(Professor Thomas F. Hofmann - TUM President)

The MCML will receive permanent funding from the German government and the state of Bavaria. As a result, regional research into AI and particularly ML will gain considerable cachet within the knowledge hub of Munich and beyond.

«The digital revolution is moving into the next phase. Machine learning is playing an increasingly prominent role – including in application: industry, mobility, the care sector. In MCML, we are pooling the AI expertise of our two universities of excellence. This is great teamwork between LMU and TUM.»
(Markus Blume - Bavarian State Minister of Science and the Arts)

Research on artificial intelligence firmly anchored in Germany.

In his lecture "Artificial Intelligence - A look towards future developments" our Director Thomas Seidl will talk about the latest developments in artificial intelligence. Besides reflecting on the terminology and limitations of artificial intelligence, the lecture will present selected practical issues related to machine learning, particularly around the preparation and analysis of data necessary for automated learning.

Congrats to our Junior Member Max Berrendorf and our PI Volker Tresp for receiving the Best Paper Award in the Research Track at the International Semantic Web Conference (ISWC 2021) for their work "Improving Inductive Link Prediction Using Hyper-Relational Facts".

Congrats from us!

The 1st AutoML Fall School will cover core topics of AutoML, covering basics, state-of-the-art approaches and hands-on sessions. Enthusiastic AutoML experts will present their diverse views on AutoML to ML practitioners, developers, research engineers, researchers and students.

The 1st AutoML Fall School is co-organized by our Director Bernd Bischl and his team.

The LWDA 2021 conference will be hosted by the Munich Center for Machine Learning (MCML) from September 01, 2021 to September 03, 2021. The conference will be fully online.

We invited some of the leading ML researchers from France and Germany to this two day symposium to give a glimpse into their research, and engage in discussions on the future of machine learning and how to strengthen research collaborations in ML between France and Germany.

The symposium is co-organized by our Director Daniel Cremers and his team.

Our partner MTU Aero Engines AG in Munich is looking for you

Our partner MTU Aero Engines AG in Munich is looking for you

Our partner MTU Aero Engines AG in Munich is looking for you

Our partner MTU Aero Engines AG in Munich is looking for you

Technology pioneer: The Munich Center for Machine Learning and the Leibniz Supercomputing Center have installed a new data-processing architecture for AI research.

This workshop will have the purpose to share insights, discuss emerging topics and a strategic roadmap for the field under the title 'Open Challenges and Future Directions of NLP'.

Researchers are obliged to conform to rules of Good Scientific Practice. To raise awareness about these questions, life science graduate programs in the greater Munich area.

The Graduate Center of LMU and MCML have joined forces to organize a campus-wide educational event.

LMU Munich joins forces with University of Warwick under the DSSGx UK chapter of the DSSG Foundation supported by the Alan Turing Institute to organise 2021 DSSGx Summer Projects.

"Give me your data and we’ll find patterns," says MCML Director Thomas Seidl. Decoding such patterns reveals secrets hidden in everything from bone fossils to ancient languages. And as academia and industry collect more and more data, there is growing interest in using artificial intelligence (AI) techniques to analyse them.

Read more about LMU's focus areas in AI research in this article.

LMU is now partner of the Marie Curie Innovative Training Network entitled "Machine Learning Frontiers in Precision Medicine" (MLFP).

MLFP brings together leading European research institutes in machine learning and statistical genetics, both from the private and public sector, to train 14 early stage researchers.

These scientists will apply machine learning methods to health data. The goal is to reveal new insights into disease mechanisms and therapy outcomes and to exploit the findings for precision medicine, which hopes to offer personalized preventive care and therapy selection for each patient.

MCML PI Volker Tresp, Yunpu Ma and Zhen Han review recently developed learning-based algorithms for temporal knowledge graphs completion and forecasting. Knowledge graphs, also known as episodic or time-dependent knowledge graphs are large-scale event databases that describe temporally evolving multi-relational data.

The talk recalls commonalities like model-based supervised and unsupervised learning approaches, and it also sheds light to peculiarities in process data and process mining tasks as seen from a data mining perspective.

NCML Junior Member Moritz Herrmann receives the Bernd-Streitberg Award of the German Biometric Society (IBS-DR) for his master thesis "Large-scale benchmark study of prediction methods using multi-omics data" supervised by our PI Anne-Laure Boulesteix.

The workshop includes presentations on Spatial and Temporal Machine Learning & Computer Vision, Learning on Graphs and Networks & Representation Learning, and Automatic and Explainable Modeling & Computational Models for Large-Sclae ML.

As a member of the PyKEEN community project, we are happy to announce PyKEEN 1.0 – PyKEEN is a software package to train and evaluate knowledge graph embedding models.

This enables us to intensively expand the existing research groups and extend the research areas by three additional project groups.