18.12.2025
MCML Thinkathon
A Short Recap
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 …
18.12.2025
“See, Don’t Assume”: Revealing and Reducing Gender Bias in AI
MCML Research Insight - With Leander Girrbach, Yiran Huang, Stephan Alaniz and Zeynep Akata
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 …
16.12.2025
Fabian Theis Featured in Handelsblatt on the Future of AI in Precision Medicine
Artificial Intelligence and Biomedical Research
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 …
16.12.2025
Gitta Kutyniok Featured in VDI Nachrichten on AI Ethics
Defining Measurable Criteria for Responsible AI Agents
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 …
16.12.2025
Hinrich Schütze Featured in WirtschaftsWoche on Innovative AI Approaches
Advancing Language Models With Memory-Efficient and Context-Aware AI
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 …
16.12.2025
Xiaoxiang Zhu Featured in Focus Online on Global 3D Building Atlas
AI-Driven 3D Mapping Reveals Urbanization, Infrastructure, and Social Insights
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 …
11.12.2025
Digdeep Podcast: Will AI Be the New Colleague, Prof. Sabine Pfeiffer?
News From the Digital World by MCML PI Frauke Kreuter and Christof Horn
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. …
11.12.2025
From Sitting Dog to Standing: A New Way to Morph 3D Shapes
MCML Research Insight - With Lu Sang and Daniel Cremers
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”.
©BAdW / Karl Neunert
08.12.2025
Tom Sterkenburg Wins Karl-Heinz Hoffmann Prize of the Bavarian Academy of Sciences
Honored for His Work at the Intersection of Philosophy, Statistics, and Machine Learning
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 …
08.12.2025
Björn Ommer Guest on ZDF’s Terra X
Image-Generative AI and Visual Understanding
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, …
08.12.2025
Björn Schuller Guest on ZDF’s Terra X
Emotion Recognition and Empathic AI
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 …
08.12.2025
MCML General Assembly
A Short Recap
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.
©Juli Eberle / TUM / ediundsepp Gestaltungsgesellschaft
04.12.2025
World’s First Complete 3D Model of All Buildings Released
TUM News
A research team led by MCML PI Xiaoxiang Zhu has released the GlobalBuildingAtlas, the first global high-resolution 3D model of all buildings on Earth. Based on 2019 satellite data, the atlas includes more than 2.75 billion buildings worldwide. It offers a ground-breaking spatial resolution of 3×3 meters, providing up to 30 times more detail than …
04.12.2025
When to Say “I’m Not Sure”: Making Language Models More Self-Aware
MCML Research Insight - With Yawei Li, David Rügamer, Bernd Bischl, and Mina Rezaei
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” …