Di He

Quotient-Space Diffusion Models

In-Place Test-Time Training

Diagnosing and Improving Diffusion Models by Estimating the Optimal Loss Value

From Five Dimensions to Many: Large Language Models as Precise and Interpretable Psychological Profilers

Online Pseudo-Zeroth-Order Training of Neuromorphic Spiking Neural Networks

Understanding vs. Generation: Navigating Optimization Dilemma in Multimodal Models

MIRAGE: MULTI-HOP INTERLEAVED REASONING AND RETRIEVAL-GROUNDED EVIDENCE

Playing with Transformer at 30+ FPS via Next-Frame Diffusion

S-ATM: Self-Boosting Visual Reasoning via Adaptive Token Merging

Luminark: Training-free, Reliable Watermarking for General Vision Generative Models

Theoretical Benefit and Limitation of Diffusion Language Model

DPO Meets PPO: Reinforced Token Optimization for RLHF

Let the Code LLM Edit Itself When You Edit the Code