Rui Zhao

4 papers Tencent

• CoC-VLA: Delving into Adversarial Domain Transfer for Explainable Autonomous Driving via Chain-of-Causality Visual-Language-Action Model
• DiffE2E: Rethinking End-to-End Driving with a Hybrid Diffusion-Regression-Classification Policy
• STAR: Efficient Preference-based Reinforcement Learning via Dual Regularization
• UniLumos: Fast and Unified Image and Video Relighting with Physics-Plausible Feedback