Schedule

Introduction and opening remarks

Invited Talk 1 (Yingbin Liang)

Contributed Talk 1 : Universal Sharpness Dynamics in Neural Network Training: Fixed Point Analysis, Edge of Stability, and Route to Chaos

Poster Session 1 and Coffee Break 

Contributed Talk 2: Catapults in SGD: spikes in the training loss and their impact on generalization through feature learning

Contributed Talk 3: Interpreting Grokked Transformers in Complex Modular Arithmetic

Contributed Talk 4: What Can Transformer Learn with Varying Depth? Case Studies on Sequence Learning Tasks

Contributed Talk 5: Promises and Pitfalls of Generative Masked Language Modeling: Theoretical Framework and Practical Guidelines

Poster Session 2 and Lunch Break