MCML Munich AI Day

A Short Recap

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.

Panel about the future of AI. From left to right: Moderator Larissa Holzki (Handelsblatt), MCML PI Barbara Plank, MCML Director Daniel Cremers, Andreas Liebl (appliedAI Initiative GmbH), StM Markus Blume (StMWK), and Tina Kluewer (BMBF)

How can we make advances in the foundations of AI? How can we utilize AI in applications relevant to society’s needs, from medicine to autonomous driving? And in doing so, how can we ensure that AI systems are safe and reliable? These were questions our keynote speakers and panelists addressed.

The event highlighted the broad spectrum of research conducted at the MCML in the international and application-driven context and inspired conversations between researchers and industry representatives.

Michal Irani

In her keynote, Michal Irani from the Weizmann Institute of Science discussed

(1) How to reconstruct images that a person has seen directly from their fMRI brain recordings.

(2) How to reconstruct the training data that a deep network has been trained on directly from the parameters of the network.

Mihaela van der Schaar

Mihaela van der Schaar from the University of Cambridge spoke on how in today’s rapidly evolving world, technology is profoundly transforming healthcare. She explored five key elements driving this change, focusing on how Artificial Intelligence can shape the future of medicine and healthcare.

MCML PI Hinrich Schütze

MCML PI Hinrich Schütze from Ludwig-Maximilians-Universität München gave innovative insights into his research in Natural Language Processing. He showed how to train LLMs to become successful autodidacts. He discussed the general principles of successful “self-training” and how to apply them to the case of improving an LLM’s capability to produce long form content. This approach is promising for many skills and domains for which training data are sparse and too expensive to generate manually.

MCML PI Stefanie Jegelka

MCML PI Humboldt-Professor Stefanie Jegelka from Technische Universität München showed how data in form of graphs are becoming more and more successful in many applications, and discussed some recent research questions and results about modeling, stability and transferability of such models.

Booth by baiosphere

Since one of the core missions of the MCML is to strengthen relations between academia and industry, we also had invited companies. Booths by DeepScenario, HP, and SE3 Labs as well as from the baiosphere, the Bavarian AI network, attracted great interest. Moreover, Federico Tombari from Google gave a keynote on 3D representations.

Poster session by MCML Junior Members

In addition to the talks, posters by junior members of the MCML highlighted the broad range of research at the MCML and gave an insight into the current research and future developments. Attendees had the opportunity to engage directly with the junior researchers, discuss their methodologies, and explore potential collaborations.

Markus Blume, Bavarian Minister of Science and Arts

Markus Blume, Bavarian Minister of Science and Arts stated:

« With the Munich AI Day, the Munich Centre for Machine Learning offers an important platform for artificial intelligence. One thing is clear: the biggest risk with AI is not being there. We have decided: We want to shape the key technology of the future ourselves according to our values. This is the only way to achieve technological and economic sovereignty in the future. […] The MCML is our Bavarian think tank for AI - thank you very much for your important work! »

The MCML has established itself as one of the largest and most prominent AI research centers in the European ecosystem and, over the past year, has succeeded in forging ties to prominent researchers from Germany and abroad.