Best paper award for Sameer Ambekar

Congratulations! relAI student Sameer Ambekar wins the best paper award at the MICCAI Workshop on Advancing Data Solutions in Medical Imaging AI (ADSMI). 

Sameer is a PhD student at relAI, advised by the relAI Fellow Julia A. Schnabel. His research focusses on test-time adaptation and domain generalization for medical imaging. 

His award-winning paper “Selective Test-Time Adaptation for Unsupervised Anomaly Detection using Neural Implicit Representations”, co-authored with Julia A. Schnabel and Cosmin Bereca, presents a novel zero-shot methodology to adapt models in real time to test images from new domains using deep pre-trained features. The approach is validated on brain anomaly detection data. 

This work addresses domain shift at test-time, which Sameer explains in more detail in his recently published relAI blog post. In the post, you can also learn about the importance of handling domain shifts to make AI more reliable: https://zuseschoolrelai.de/blog/mitigating-domain-shifts/  

Congratulations on this achievement!  

Congratulations to relAI fellow Johannes Maly on winning the 2024 ACHA Charles Chui Young Researcher Best Paper Award!

This annual award honors the exceptional contributions of young researchers in the field of harmonic analysis, selected by the editors and publisher of the Journal Applied and Computational Harmonic Analysis (ACHA)

Johannes received this recognition for his paper, “Robust Sensing of Low-Rank Matrices with Non-Orthogonal Sparse Decomposition,” published in November 2023 (Vol. 67, ACHA). His research focuses on understanding how to leverage intrinsic models or data structures and how the loss of information caused by digital processing (quantization) affects theoretical results. This is highly relevant for resource-aware and reliable AI

We are excited to announce that relAI fellow Daniel Rückert has been awarded the prestigious MICCAI Enduring Impact Award 2024

Daniel Rückert holds the Alexander von Humboldt Professorship for AI in Medicine and Healthcare at TUM. His research focuses on developing advanced algorithms for biomedical image analysis, including segmentation, registration, and deep learning methods to support reliable AI in healthcare applications. 

The MICCAI Society (Medical Image Computing and Computer-Assisted Intervention) recognized Daniel Rückert for his outstanding, lasting contributions to the field. His work has significantly advanced the application of AI in biomedical imaging, helping create reliable, clinically impactful solutions. 

Congratulations on this achievement!  

We are excited to congratulate our relAI fellow, Prof. Björn Ommer, on receiving the prestigious 2024 German AI Innovation Award!   

This recognition honors his pioneering work on Stable Diffusion, a generative AI model that has democratized image generation. Stable Diffusion enables users to create high-quality images from text descriptions, making advanced image generation accessible through open-source software without the need for expensive hardware. This innovation exemplifies our mission at relAI, which focuses on advancing reliable AI systems. 

The German AI Award, presented by WELT, recognizes innovations that drive the future of AI both in theory and application. Björn Ommer’s recognition reflects his significant role in shaping the evolution of reliable and transparent AI systems. 

Our heartfelt congratulations! 

Figure: © Fabian Helmich; Graphic generated with stable diffusion. 

relAI is proud to announce the addition of five exceptional fellows to the school. Enkelejda Kasneci (TUM), Gjergji Kasneci (TUM), Björn Ommer (LMU), Tom Sterkenburg (LMU), and Abdalla Swikir (TUM) have each made significant contributions to their respective fields and bring a wealth of knowledge and experience to our family.

The research topics of the new fellows tackle essential aspects of the field of reliable AI. Their work ranges from the study of human-computer interaction and semantic scene understanding to the study of fairness and inductive bias in machine learning as well as safe learning in robotics. Their engagement in the school's research and educational activities will contribute to the reliable application of AI in real-life scenarios, such as improving user experience in digital interfaces and enhancing the safety of autonomous systems.

Enkelejda Kasneci is a Distinguished Professor (“Liesel Beckmann Distinguished Professorship”) for Human-Centered Technologies for Learning at the TUM School of Social Sciences & Technology. Her research focuses on Human-Computer Interaction and developing AI systems that sense and infer the user’s cognitive state, expertise, actions, and intentions based on multimodal data.

Gjergji Kasneci holds the Chair for Responsible Data Science at TUM School of Computation, Information & Technology. His research focuses on transparency, robustness, bias, and fairness in machine learning algorithms, incorporating ethical, legal, and societal considerations.

Abdalla Swikir is a Senior Scientist and Teaching Coordinator at the TUM Munich Institute of Robotics and Machine Intelligence (MIRMI). His research in safe learning for robotic control and autonomous systems targets the enhancement of reliability and safety, ensuring these technologies can effectively function in dynamic and critical environments.

Björn Ommer is Head of the LMU Computer Vision & Learning Group. His research interests include semantic scene understanding and retrieval, generative AI and visual synthesis, explainable AI, and self-supervised metric and representation learning. Moreover, he is applying this basic research in interdisciplinary projects within neuroscience and the digital humanities.

Tom Sterkenburg is an Emmy Noether junior research group leader at the Munich Center for Mathematical Philosophy at LMU Munich. His Emmy Noether project, “From Bias to Knowledge: The Epistemology of Machine Learning", is concerned with clarifying the fundamental concept of inductive bias in machine learning.

relAI Fellow Pramod Bhatotia wins the 2024 Rising Star in Dependability Award: 

 “For his impressive track-record and contributions in the field of dependable systems, including multiple publications in highly regarded venues, and influence on practical dependable systems.” 

Pramod Bhatotia is a professor for Systems Research at TU Munich and fellow at relAI. He received the award at the 54th Annual IEEE/IFIP International Conference on Dependable Systems and Networks, which took place last week in Brisbane, Australia.  

The award aims to recognize a junior researcher, “who demonstrates outstanding potential for creative ideas and innovative research in the field of dependable and resilient computer systems and networks”.  

More information: https://dsn2024uq.github.io/awards_rs.html 

Our congratulations! 

Image Copyright (c): Thomas Abé/Studienstiftung

Congratulations to relAI student Maria Matveev

The German National Scholarship Foundation (Studienstiftung) has awarded Maria the Civic Engagement Award 2024 for her exceptional volunteering work with Lern-Fair. Maria co-founded and chairs Lern-Fair e.V., a non-profit organization dedicated to providing free educational opportunities for underprivileged pupils. Since the start of the online platform in 2020 during the Covid pandemic, more than 15.000 pupils were supported by free tutoring or group courses. 

Maria is a relAI PhD student at the chair for Mathematical Foundations of Artificial Intelligence at LMU and the Munich Center for Machine Learning. Her PhD research, advised by the relAI director Gitta Kutyniok, focuses on the mathematical description and understanding of training dynamics related to generalization, a crucial factor for ensuring the reliability of neural networks. 

Learn more about Marias volunteer work in the video portrait (in German): https://youtu.be/EUZdm--sqmc?feature=shared 

We are delighted to welcome four exceptional fellows to relAI: Nassir Navab, Solveig Vieluf, Tobias Lasser and Jochen Kuhn.  

Each of them brings their own expertise and insights that will further enrich our research agenda, educational offers, and scientific community. They are dedicated to making significant contributions to the advancement of reliable AI, particularly in Medicine & Healthcare and AI in Education areas.  

Nassir Navab is a full professor and director of the Laboratories for Computer Aided Medical Procedures at TUM and adjunct professor at John Hopkins University. One focus of his research is AI assisted Surgery, where reliable methods are a key requirement for both clinicians and patients.  

Solveig Vieluf is a professor of AI-based telemonitoring in the field of cardiology at LMU. Previously, she has also worked on epilepsy and aging research. She uses methods from explainability to explore influence factors on model performance. 

The research of Tobias Lasser is focused on computational imaging and inverse problems in medicine and healthcare. In his work on clinical decision support using AI, he works on prioritization of critical cases for treatment.  

Jochen Kuhn works on the intersection of AI and education, in particular on the use of these future technologies to foster learning and teaching in STEM disciplines. He is a professor of Physics education at LMU. Reliability is important in his research, particularly the role of bias and inaccurate information from AI chatbots on learning and teaching.  

Join us in welcoming these four to the relAI community! 

We are proud to announce that the German Radiological Society (Deutsche Röntgengesellschaft) has awarded the Alfred Breit Prize 2024 to our relAI fellow Prof. Julia Schnabel. The prize honors outstanding work and developments in the field of radiological research that have significantly contributed to progress in cancer therapy.

Julia Schnabel is Professor for Computational Imaging and AI in Medicine at the Technical University of Munich TUM (Liesel Beckmann Distinguished Professorship), and Director at the Institute for Machine Learning in Biomedical Imaging at Helmholtz Munich (Helmholtz Distinguished Professorship). Since 2015, she has also been Professor of Computational Imaging at King's College London.

Prof. Schnabel works in the field of medical image processing and machine learning. Her research focuses on the areas of intelligent imaging up to clinical evaluation, including complex motion modeling, image reconstruction, quality assurance, segmentation, and classification applied to multimodal, quantitative, and dynamic imaging.

Congratulations!

Are you interested in frontier AI systems, their astonishing capabilities and risks for humanity? Then join us for a thought-provoking deep dive and exclusive OpenAI Live Q&A on AI safety. 

  • Date: Wednesday, May 8th, 2024 | 19:00 – 20:30 
  • Location: Room B006, Department of Mathematics (Theresienstr. 39) or online 
  • Language: English 

Agenda: 

  • 19:00 – 19:05: Doors open 
  • 19:05 – 19:30: Introduction to AI Safety 
  • 19:30 – 20:15: Presentation & Live Q&A with OpenAI researcher Jan H. Kirchner, co-author of weak-to-strong generalization paper 
  • 20:15 – 20:30: Closing talk – What can we do? 
  • 20:30 – onward: Optional socializing and small group discussions with free drinks and snacks. 

Please register on the following webpage and prepare your questions!