Deep Learning Theory Workshop 2025

Description

The Deep Learning Theory Workshop 2025 is organized by the Deep Learning Theory Team at RIKEN AIP, and aims to foster in-depth discussions and collaborations at the intersection of deep learning practice and theory.

This workshop brings together leading researchers from Japan and around the world, including both academic experts and members of our team, to discuss recent advances, open problems, and emerging trends in the theoretical understanding of deep learning.

While deep learning continues to achieve remarkable success across a wide range of applications, the theoretical foundations behind these breakthroughs are still being actively explored. Questions around generalization, optimization, representation, and robustness remain central to developing a principled understanding of modern deep learning systems.

We invite contributions and participation from researchers in learning theory, machine learning, statistics, information theory, and related fields, with the goal of building bridges between theoretical analysis and practical advances in deep learning.

Schedule

Day 1
10:00–10:50 Taiji Suzuki

11:00–11:50 Murat Erdogdu

12:00–13:30 Lunch time

13:30–14:20 Sho Sonoda

14:30–15:20 Shinho Chewi

15:30–16:20 Mahdi Soltanolkotabi

16:30–17:30 Denny Wu

17:30–18:30 Poster session
18:30– Dinner time

Day 2
10:00–10:50 Atsushi Nitanda

11:00–11:50 Yuan Cao

12:00–13:30 Lunch time
13:30–14:20 Wei Huang

14:30–15:20 Fanghui Liu

15:30–16:20 Matus Telgarsky

16:30–17:20 Takashi Furuya

18:00– Dinner time

Day 3
10:00–10:50 Han Bao

11:00–11:50 Nikhil Ghosh

11:50–12:00 Closing remark