Game-Theoretic Diffusion Models

11.06.2024 09:50 - 10:35

Ya – Ping Hsieh (ETH Zürich)

AI systems hold great potential in advancing both industrial and scientific domains, with a notable impact on areas such as AI-assisted drug discovery. Despite their promise, conventional AI models frequently function within static and single-agent decision-making frameworks. In this presentation, I aim to explore the inherent challenges posed by these current frameworks and present algorithmic solutions designed to overcome their limitations. Following this, I will introduce a research project dedicated to integrating multi-agent paradigms and dynamic modeling. Our approach seeks to augment the practical efficacy of upcoming AI systems by bridging the gaps between established theoretical frameworks and the complexities of real-world scenarios.


Fakultät für Mathematik, Dekan Radu Ioan Boţ

Hörsaal 12, 2. OG