November 11, 2025 13:25
[RIKEN AIP PRM Series] Deep Learning Theory Workshop 2025 Day2 (Aug. 6, 2025) thumbnails

Description

Deep Learning Theory Workshop 2025 Day2 (Aug. 6, 2025)
＊You can start each talk by clicking the time (h:m:s) before the speaker’s name.
00:00:06 Atsushi Nitanda “Propagation of Chaos for Mean-Field Langevin Dynamics”
00:54:32 Yuan Cao “Token Selection in Self-Attention Mechanisms: Case Studies and Theoretical Insights”
01:47:37 Wei Huang “Understanding Benign Overfitting in Vision Transformers”
02:36:54 Fanghui Liu “Bridge theory to practice: One-step full gradient can suffices for low-rank fine-tuning, provably and efficiently”
03:25:45 Matus Telgarsky “Open problems in Optimization path following and Transformer Approximation”
04:07:59 Takashi Furuya “Transformers are universal in-context learners”

[Day 1] Click the following URL

Event page URL