🎉The 8th edition of DataFest Germany will be held at Ludwig-Maximilians-Universität in Munich from 28 March to 30 March 2025. relAI is proud to support the event organization again this year. Additionally, a team of relAI students will participate in this exciting competition and networking opportunity.
The event is an annual data-driven competition, commonly referred to as a “hackathon,” that alternates between Mannheim and Munich. It is organized in collaboration with partners from industry and research institutions.
Datafest Germany is a celebration that follows upon the model DataFest™, organized by the American Statistical Association. The world's first DataFest took place at the University of California in 2011. Since then, many universities took up the DataFest format.
Title:Â Foundational Methods for Foundation Models for Scientific Machine Learning
Date and Time: March 26, 2025 14:00 CET
Location: Lecture Hall W201, Professor-Huber-Platz 2, LMU Munich, 80539 Munich (Metro U3/U6 Universität, Exit B) LMU Room Finder
Abstract
The remarkable successes of ChatGPT in natural language processing (NLP) and related developments in computer vision (CV) motivate the question of what foundation models would look like and what new advances they would enable, when built on the rich, diverse, multimodal data that are available from large-scale experimental and simulational data in scientific computing (SC), broadly defined. Such models could provide a robust and principled foundation for scientific machine learning (SciML), going well beyond simply using ML tools developed for internet and social media applications to help solve future scientific problems. Prof. Mahoney will describe recent work demonstrating the potential of the "pre-train and fine-tune" paradigm, widely-used in CV and NLP, for SciML problems, demonstrating a clear path towards building SciML foundation models; as well as recent work highlighting multiple "failure modes" that arise when trying to interface data-driven ML methodologies with domain-driven SC methodologies, demonstrating clear obstacles to traversing that path successfully. Prof. Mahoney will also describe initial work on developing novel methods to address several of these challenges, as well as their implementations at scale, a general solution to which will be needed to build robust and reliable SciML models consisting of millions or billions or trillions of parameters.
Bio of the speaker
Michael W. Mahoney is at the University of California at Berkeley in the Department of Statistics and at the International Computer Science Institute (ICSI). He is also an Amazon Scholar as well as head of the Machine Learning and Analytics Group at the Lawrence Berkeley National Laboratory. He works on algorithmic and statistical aspects of modern large-scale data analysis. Much of his recent research has focused on large-scale machine learning, including randomized matrix algorithms and randomized numerical linear algebra, scientific machine learning, scalable stochastic optimization, geometric network analysis tools for structure extraction in large informatics graphs, scalable implicit regularization methods, computational methods for neural network analysis, physics informed machine learning, and applications in genetics, astronomy, medical imaging, social network analysis, and internet data analysis. He received his PhD from Yale University with a dissertation in computational statistical mechanics, and he has worked and taught at Yale University in the mathematics department, at Yahoo Research, and at Stanford University in the mathematics department. Among other things, he was on the national advisory committee of the Statistical and Applied Mathematical Sciences Institute (SAMSI), he was on the National Research Council's Committee on the Analysis of Massive Data, he co-organized the Simons Institute's fall 2013 and 2018 programs on the foundations of data science, he ran the Park City Mathematics Institute's 2016 PCMI Summer Session on The Mathematics of Data, he ran the biennial MMDS Workshops on Algorithms for Modern Massive Data Sets, and he was the Director of the NSF/TRIPODS-funded FODA (Foundations of Data Analysis) Institute at UC Berkeley. More information is available at https://www.stat.berkeley.edu/~mmahoney/ .
This event is open to everyone; registration is not required.
Prof. Starck will speak about how inverse problems in astrophysics, such as image reconstruction or gravitational lensing data analysis, have traditionally relied on sparsity-based techniques to recover underlying physical structures from incomplete or noisy data. Deep learning methods are now replacing these classical approaches, offering unprecedented performance gains in accuracy and efficiency. Despite their success, deep learning methods introduce new challenges, including interpretability, generalization across diverse astrophysical scenarios, and robustness to observational biases. In this talk, the speaker will explore the transition from sparsity-driven methods to deep learning-based solutions, highlighting both the opportunities and pitfalls of this paradigm shift. Prof. Starck will discuss recent developments, applications to astrophysical data, and future directions for addressing the emerging challenges in this rapidly evolving field.
To read more information about the event and the speaker, visit this weblink.
relAI is a co-organiser of the Munich AI lectures. Find more info on this and other upcoming events on the Munich AI lectures home page.
/ mc
This month, a team of 13 talented master and PhD students from our graduate school in reliable AI (relAI) showcased their quantitative skills and teamwork in an exciting estimation competition. The participants had 30 minutes to work on 13 estimation challenges, such as "What is the average discharge of the Isar when it meets the Donau in m^3/s?"
The spirit of competition and learning was truly inspiring. Check out the photo of our team, proudly representing relAI.
/ mc
We are pleased to share that on January 29, 2025, relAI will join the opening ceremony of our partner AI-HUB@LMU to celebrate its founding. AI-HUB@LMU is a platform that, for the first time, unites all 18 faculties of Ludwig-Maximilians-Universität MĂĽnchen as a joint scientific community and aims to advance research, teaching, and transfer in artificial intelligence and data science at LMU. Â
As part of our commitment to fostering collaboration and innovation, relAI supports the organization of this significant event. The event will be honoured by inaugural remarks from representatives of the university and government. All 18 faculties will then present their highlights in AI and data science in keynote talks, panel discussions, pitch talks, and presentations. Check this link for the full program: https://www.ai-news.lmu.de/grand-opening-ai-hub.
/ mc
Multi-Head Attention has become ubiquitous in modern machine learning architectures, but how much efficiency can still be gained? This question was the focus of Dr. Maximilian Baust’s talk, "Beyond Transformers: Why Beating Multi-Head Attention is Hard."
In his presentation, Dr. Baust explored potential solutions for improving efficiency, ranging from implementation strategies and algorithmic modifications to new architectures, including spiking neural networks.
Dr. Maximilian Baust serves as Director of Solution Architecture Industries EMEA at NVIDIA and is also an industry mentor for one of relAI’s PhD students.
We extend our gratitude to Dr. Baust for sharing his insights and to our director, Gitta Kutyniok, for inviting him to relAI.
/ mc
The recent relAI Collab Accelerator Workshop brought together researchers to share their work, explore new ideas, and identify potential collaborations. Here's a brief overview of the event:
The day began with the participants pitching their research topic from 9:00 to 11:00, followed by a coffee break until 11:15. After the break, participants engaged in one-to-one sessions until 14:00, followed by lunch, discussion, and feedback.
Participants contributed diverse topics in the field of Machine Learning and Artificial Intelligence. Mohamed Amine Ketata discussed Generative AI for Graphs, Max Beier presented on Learning Operator of Dynamical Systems, Richard Schwank explored Robust Aggregation through the Geometric Median, and Yurou Liang delved into Differentiable Learning for Causal Discovery.
The workshop was fertile ground for generating new research ideas and possible collaborations. During the one-to-one discussions, participants identified several projects for cooperation, such as principled modifications of loss functions to enhance robustness against outlier data rows.
Participants gained new insights into their research during the event. For example, one participant discovered a probabilistic approach to their forecasting issue without relying on a model. Another learned about structure learning as it applies to tabular data, which provided a temporal interpretation of the data. One researcher was challenged about the convexity of their problem. Discussions highlighted intriguing applications of median aggregation techniques to abstract spaces, connecting concentration inequalities with uncertainty quantification.
The relAI Collab Accelerator Workshop was an enriching experience, offering a platform for researchers to connect, share insights, and pave the way for future collaborations. The feedback during this first iteration will help refine the format and make it even more engaging. We are looking forward to the next iteration!
relAI thanks Max Beier and Richard Schwank for their initiative and the organization of the event.
/ mc
We are happy to invite you to the upcoming Munich AI Lecturefeaturing two distinguished researchers Prof. Holger Hoos from RWTH Aachen University and Prof. Franca Hoffmann from California Institute of Technology. The lecture is organized by the Chair of Mathematics of Information Processing with support by MCML.
In the first talk, “Dynamics of Strategic Agents and Algorithms as PDEs“, Prof. Hoffmann will speak about dynamics of interactions between algorithms and a population.
In the second talk, “Learning, reasoning and optimisation: Adversarial robustness of neural networks”, Prof. Hoos will discuss robustness of neural networks and its verification.
Event Details:
Speakers: Prof. Dr. Holger Hoos and Prof. Dr. Franca Hoffmann
Date and Time: December 17, 2024, 16:00 CET (16:00-17:00 talk by Prof. Hoffmann and 17:30-18:30 talk by Prof. Hoos – We will have a small break with coffee/tea and snacks in between)
Franca Hoffmann obtained her master’s in mathematics from Imperial College London (UK) and holds a PhD from the Cambridge Centre for Analysis at University of Cambridge (UK). She held the position of von Kármán instructor at Caltech from 2017 to 2020, then joined University of Bonn (Germany) as Bonn Junior Professor and Quantum Leap Africa in Kigali, Rwanda (African Institute for Mathematical Sciences) as AIMS-Carnegie ResearchChair in Data Science, before arriving at the California Institute of Technology as Assistant Professor in 2022.
Bio Holger Hoos
Holger H. Hoos holds an Alexander von Humboldt professorship in AI at RWTH Aachen University (Germany), where he also leads the AI Center, as well as a professorship in machine learning at Universiteit Leiden (the Netherlands) and an adjunct professorship in computer science at the University of British Columbia (Canada). He is a Fellow of the Association of Computing Machinery (ACM), the Association for the Advancement of Artificial Intelligence (AAAI) and the European AI Association (EurAI), past president of the Canadian Association for Artificial Intelligence, former editor-in-chief of the Journal of Artificial Intelligence Research (JAIR) and chair of the board of CLAIRE, an organization that seeks to strengthen European excellence in AI research and innovation (claire-ai.org).
relAI is a co-organiser of the Munich AI lectures. Find more info on these other upcoming events on the Munich AI lectures home page.
/ mc
Save the date!
We are pleased to invite you to the upcoming Munich AI Lecture featuring renowned researcher Prof. Dr. Helmut Bölcskei hosted by our relAI director Prof. Dr. Gitta Kutyniok, Bavarian AI Chair for Mathematical Foundations of Artificial Intelligence at LMU.
Deep neural networks have led to breakthrough results in numerous practical machine learning tasks. In the lecture “The Mathematical Universe behind Deep Neural Networks,” Prof. Dr. Bölcskei will take us on an exciting journey through the mathematical universe behind these practical successes, elucidating the theoretical underpinnings of deep neural networks in functional analysis, harmonic analysis, complex analysis, approximation theory, dynamical systems, Kolmogorov complexity, optimal transport, fractal geometry, mathematical logic, and automata theory.
Helmut Bölcskei is a Professor of Mathematical Information Science at ETH Zurich and has been a Principal Investigator at the Lagrange Mathematics and Computing Research Center in Paris since 2021. After earning his degrees from Vienna University, he completed a postdoctoral fellowship at Stanford University and has held visiting researcher positions at many leading institutions. In addition to his academic achievements, he is also a successful entrepreneur. Professor Bölcskei has received numerous awards and prestigious fellowships and continues to serve as Editor-in Chief for some of the field's most distinguished journals.
From 28 to 30 October, the second annual meeting of the three DAAD Zuse Schools of Excellence in Artificial Intelligence, ELIZA (Darmstadt), SECAI (Dresden), und relAI (Munich) took place in Munich.
The meeting kicked off on the 28th with a work-meeting of the Zuse Schools directors and coordination teams, the chairman of the Zuse Schools Advisory Board, as well as representatives of BMBF and DAAD. The main event, which followed on the 29th, celebrated the three Zuse Schools’ second anniversary with a rich program featuring welcome speeches, academic and industry keynotes, talks by Zuse School students, and a stimulating panel discussion.
The celebration on the 29th was opened by inspiring welcome addresses from Prof. Dr. Gerhard Kramer, Senior Vice President for Research and Innovation, TUM, Prof. Dr. Hans van Ess, Vice President for Research, LMU, Dr. Kai Sicks, DAAD Secretary General, MinR’in Dr. Lisette Andreae, Head of Unit European Higher Education Area, Internationalization, BMBF and MinDir Dr. Rolf-Dieter Jungk, Bavarian State Ministry for Science and the Arts (StMWK). relAI Fellow Prof. Björn Ommer delivered a brilliant keynote talk on the latest research on Generative AI. The industry keynote by Dr. Ahmed Sayed, Head of EMEA Emerging Technologies, AWS, provided an insightful overview of the industry applications of AI. The four talks from students of the three Zuse Schools showcased the excellent research carried out by the schools. The program was rounded off with an engaging panel discussion about the topic “Reliability in times of generative AI”, featuring a diverse range of perspectives and voices.
The event concluded on the 30th with a Dialogue session, “My academic and professional future as a Zuse Schools graduate in Germany – chances and challenges” and the Advisory Board Meeting.
We are thankful to all the speakers and contributors who made the meeting a success, and to BMBF and DAAD for their support of the Zuse Schools.
Photo on the right side: DAAD / Siegfried Michael Wagner
