Research Blog
31.07.2025
From Vulnerable to Verified: Exact Certificates Shield Models From Label‑Flipping
MCML Research Insight - With Lukas Gosch, Stephan Günnemann and Debarghya Ghoshdastidar
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, …
30.07.2025
Tracking Our Changing Planet From Space - With Xiaoxiang Zhu
Research Film
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 …
29.07.2025
AI for Enhanced Eye Diagnostics - With Researcher Lucie Huang
Research Film
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 at TUM, 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 …
24.07.2025
SceneDINO: How AI Learns to See and Understand Images in 3D–Without Human Labels
MCML Research Insight - With Christoph Reich, Felix Wimbauer, and Daniel Cremers
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 …
23.07.2025
How Reliable Are Machine Learning Methods? With Anne-Laure Boulesteix and Milena Wünsch
Research Film
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?
16.07.2025
AI-Powered Cortical Mapping for Neurodegenerative Disease Diagnoses - With Christian Wachinger
Research Film
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.
10.07.2025
Beyond Prediction: How Causal AI Enables Better Decision-Making - With Stefan Feuerriegel
Research Film
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 …
09.07.2025
Capturing Complexity in Surgical Environments
MCML Research Insight - With Ege Özsoy, Chantal Pellegrini, Felix Tristram, Kun Yuan, David Bani-Harouni, Matthias Keicher, Benjamin Busam and Nassir Navab
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 …
06.07.2025
How Neural Networks Are Changing Medical Imaging – With Reinhard Heckel
Research Film
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.
26.06.2025
What Words Reveal: Analyzing Language in the Trump–Harris 2024 Debate
MCML Research Insight - With Philipp Wicke
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.
25.06.2025
When Clinical Expertise Meets AI Innovation – With Michael Ingrisch
Research Film
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.
23.06.2025
Autonomous Driving: From Infinite Possibilities to Safe Decisions— With Matthias Althoff
Research Film
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.
20.06.2025
Zooming in on Moments: ReVisionLLM for Long-Form Video Understanding
MCML Research Insight - With Tanveer Hannan and Thomas Seidl
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 …
12.06.2025
Why Causal Reasoning Is Crucial for Reliable AI Decisions
MCML Research Insight - With Christoph Kern, Unai Fischer-Abaigar, Jonas Schweisthal, Dennis Frauen, Stefan Feuerriegel and Frauke Kreuter
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?
11.06.2025
Better Data, Smarter AI: Why Quality Matters – With Frauke Kreuter
Research Film
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.
30.04.2025
Who Spreads Hate?
MCML Research Insight – With Dominique Geissler, Abdurahman Maarouf, and Stefan Feuerriegel
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 …
29.04.2025
How Certain Is AI? An Introduction to Bayesian Deep Learning
Researcher in Focus: Emanuel Sommer
MCML Junior Member Emanuel Sommer is a PhD-student at the Munich Uncertainty Quantification AI Lab at LMU Munich supervised by David Rügamer. His research focuses on Scalable and Reliable (Bayesian) Deep Learning.
10.04.2025
Text2Loc: A Smarter Way to Navigate With Words
MCML Research Insight - With Yan Xia, Zifeng Ding and Daniel Cremers
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.
03.04.2025
CUPS: Teaching AI to Understand Scenes Without Human Labels
MCML Research Insight - With Christoph Reich, Nikita Araslanov, and Daniel Cremers
What matters now
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 …
27.03.2025
Beyond the Black Box: Choosing the Right Feature Importance Method
MCML Research Insight - With Fiona Katharina Ewald, Ludwig Bothmann, Giuseppe Casalicchio and Bernd Bischl
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.
13.03.2025
ReNO: A Smarter Way to Enhance AI-Generated Images
MCML Research Insight - With Luca Eyring, Shyamgopal Karthik, Karsten Roth and Zeynep Akata
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 …