5 papers
Mitigating Compounding Error via Video Representation Regularization
Taiye Chen, Qi Zhang, Yisen Wang
Video diffusion-based world models enable long autoregressive video generation for robotics, autonomous driving and simulation tasks, yet sliding-window autoregressive inference su…
Recurrent Autoregressive Diffusion: Global Memory Meets Local Attention
Taiye Chen, Zihan Ding, Anjian Li +4
Recent advancements in video generation has shifted from bidirectional models for short videos to autoregressive ones for ultra long video generation. Previous models, which usuall…
Scalable Defense against In-the-wild Jailbreaking Attacks with Safety Context Retrieval
Taiye Chen, Zeming Wei, Ang Li +1
Large Language Models (LLMs) are known to be vulnerable to jailbreaking attacks, wherein adversaries exploit carefully engineered prompts to induce harmful or unethical responses.…
VRAG: Learning World Models for Interactive Video Generation
Taiye Chen, Xun Hu, Zihan Ding +1
Foundational world models must be both interactive and preserve spatiotemporal coherence for effective future planning with action choices. However, present models for long video g…
Mitigating Reward Over-Optimization in RLHF via Behavior-Supported Regularization
Juntao Dai, Taiye Chen, Yaodong Yang +2
Reinforcement learning from human feedback (RLHF) is an effective method for aligning large language models (LLMs) with human values. However, reward over-optimization remains an o…