Chunshu Wu

Zeros can be Informative: Masked Binary U-Net for Image Segmentation on Tensor Cores

Latent Adaptation with Masked Policy for Diffusion Language Models

DS-TG: Dynamical Systems as Accurate and Efficient Solvers for Time-Dependent Differential Equations

Diff-PIC: Revolutionizing Particle-In-Cell Nuclear Fusion Simulation with Diffusion Models

InstaTrain: Adaptive Training via Ultra-Fast Natural Annealing within Dynamical Systems

DS-LLM: Leveraging Dynamical Systems to Enhance Both Training and Inference of Large Language Models

An Expressive and Self-Adaptive Dynamical System for Efficient Function Learning

Inertial Confinement Fusion Forecasting via Large Language Models

DM-Tune: Quantizing Diffusion Models with Mixture-of-Gaussian Guided Noise Tuning

A Brain-Inspired Machine Learning Paradigm for Nature-Powered Equation Solving