Bingrui Li

Achieving low-bit Muon through subspace preservation and grid quantization

Efficient Hyperparameter Tuning via Trajectory Invariance Principle

How Does Local Landscape Geometry Evolve in Language Model Pre-Training?

On the Optimization and Generalization of Two-layer Transformers with Sign Gradient Descent

Sharpness-Aware Minimization Efficiently Selects Flatter Minima Late In Training

A Regularized Newton Method for Nonconvex Optimization with Global and Local Complexity Guarantees

When Bigger is Better: Revisiting Large-Batch Optimization in Language Model Pretraining