Separation of timescales controls feature learning and overfitting in large neural networks

CMSA Interdisciplinary Science Seminar Speaker: Sumit Sinha, Harvard University Title: Statistical Mechanical theory for spatio-temporal evolution of Intra-tumor heterogeneity in cancers: Analysis of Multiregion sequencing data (https://arxiv.org/abs/2202.10595) Abstract: Variations in characteristics from one region (sub-population) to another are commonly observed in complex systems, such as glasses and a collection of cells. Such variations are manifestations of heterogeneity, whose […]

Interdisciplinary Science Seminar Speaker: Yang P. Liu Title: Recent Advances on Maximum Flows and Minimum-Cost Flows Abstract: We survey recent advances on computing flows in graphs, culminating in an almost linear time algorithm for solving minimum-cost flow and several other problems to high accuracy on directed graphs. Along the way, we will discuss intuitions from linear […]

Interdisciplinary Science Seminar Speaker: Yue M. Lu, Harvard University Title: Exploring and Exploiting the Universality Phenomena in High-Dimensional Estimation and Learning Abstract: Universality is a fascinating high-dimensional phenomenon. It points to the existence of universal laws that govern the macroscopic behavior of wide classes of large and complex systems, despite their differences in microscopic details. The notion of […]

CMSA Interdisciplinary Science Seminar Speaker: Quanquan Liu, Northwestern University Title: Scalable Dynamic Graph Algorithms Abstract: The field of dynamic graph algorithms seeks to understand and compute statistics on real-world networks that undergo changes with time. Some of these networks could have up to millions of edge insertions and deletions per second. In light of these […]