Sheaves in Statistical Mechanics and Machine Learning
组织者
演讲者
Grégoire Sergeant-Perthuis
时间
2025年02月21日 09:50 至 10:50
地点
A3-3-201
线上
Zoom 435 529 7909
(BIMSA)
摘要
Sheaves, which are prominent objects in geometry, topology, and algebra, have emerged numerous times in engineering over the past decades, including—but not limited to—information theory, data fusion, decentralized optimization, point cloud analysis, and deep learning. After motivating the use of sheaves in engineering, particularly for interacting systems, and demonstrating their applications in data science, the speaker will present work starting from his PhD that illustrates how they naturally arise in statistical mechanics and energy-based modeling, leading to their application in computational biology.