Dear AI Community,

It is with great pleasure that we announce the successful initiation of the project "Next Generation AI Computing (gAIn)"! This collaboration between LMU and TUM in Bavaria and TU Dresden in Saxony is financially supported by the Bavarian State Ministry of Science and the Arts, as well as the Saxon State Ministry for Science, Culture, and Tourism. The goal of the project is to develop a comprehensive, mathematics-based concept for the next generation of "Green AI" systems. The focus will be on application-based AI hardware-software combinations aimed at maximizing energy efficiency, trustworthiness, and legal compliance.

To celebrate this milestone, we invite you to our opening event which will feature keynotes from Minister Blume (Bavaria) and Minister Gemkow (Saxony) and a scientific lecture by Prof. Dr. Wolfram Burgard (UTN), as well as general information about the project. You can look forward to an exciting exchange on the future of AI and its sustainable development.

Further information and the program can be found here (https://sites.google.com/view/gainopeningagenda), along with the link to register.

The talk focuses on two the 66 problems stated in my book Geometric Tomography that are arguably the most notorious in convex geometry: Bourgain’s hyperplane conjecture from 1986 (finally confirmed last year) and Mahler’s conjecture from 1939 (still open). Both have great significance, not only for convex geometry, but for mathematics as a whole. A fairly comprehensive survey of these two important questions will be attempted, including their relationship to each other and to many other open conjectures from diverse areas of mathematics.

On April 11, 2025, the Department of Mathematics will celebrate Peter Gritzmann's 70th birthday with a festive colloquium. All TUM employees and students are cordially invited to attend the lectures in Lecture Hall 2300 on the main campus of the Technical University of Munich (TUM).

A detailed program is available at https://www.cit.tum.de/en/cit/news/article/celebratory-colloquium-peter-gritzmann/

dentifying the relevant time scales is a key step in modelling complex dynamics such as earth system components. If there are important dependencies on faster time-scale dynamics when analysing slow processes, a common approach also known as Hasselmann's program [K. Hasselmann, 1976] is to model fast chaotic influences via stochastic perturbations, as this reduces computational complexity enormously compared to integrating those fast dynamics explicitly. Apart from that, SDEs also provide us with their own tools for analysis, such as Kramers' Law, which describes transitions between different equilibria of the drift. We are able to transfer this result to the limiting chaotic case and to describe the boundaries of the validity in this case. We conclude by applying these results to a low-dimensional model of the Atlantic Meridional Overturning Circulation and investigate implications on a possible tipping of this ocean current.

We provide an algorithm giving a 140/41 (<3.415)-approximation for Coflow Scheduling and a 4.36-approximation for Coflow Scheduling with release dates. This improves upon the best known 4- and respectively 5-approximations and addresses an open question posed by Agarwal, Rajakrishnan, Narayan, Agarwal, Shmoys, and Vahdat [Aga+18], Fukunaga [Fuk22], and others. We additionally show that in an asymptotic setting, the algorithm achieves a (2+ϵ)-approximation, which is essentially optimal under P≠NP. The improvements are achieved using a novel edge allocation scheme using iterated LP rounding together with a framework which enables establishing strong bounds for combinations of several edge allocation algorithms.