5 papers
Causal World Modeling for Robot Control
Lin Li, Qihang Zhang, Yiming Luo +9
This work highlights that video world modeling, alongside vision-language pre-training, establishes a fresh and independent foundation for robot learning. Intuitively, video world…
BulletTime: Decoupled Control of Time and Camera Pose for Video Generation
Yiming Wang, Qihang Zhang, Shengqu Cai +7
Emerging video diffusion models achieve high visual fidelity but fundamentally couple scene dynamics with camera motion, limiting their ability to provide precise spatial and tempo…
Uncovering Strategic Egoism Behaviors in Large Language Models
Yaoyuan Zhang, Aishan Liu, Zonghao Ying +4
Large language models (LLMs) face growing trustworthiness concerns (\eg, deception), which hinder their safe deployment in high-stakes decision-making scenarios. In this paper, we…
Test-Time Steering for Lossless Text Compression via Weighted Product of Experts
Qihang Zhang, Muchen Li, Ziao Wang +2
Lossless compression techniques are crucial in an era of rapidly growing data. Traditional universal compressors like gzip offer low computational overhead, high speed, and broad a…
Hundreds Guide Millions: Adaptive Offline Reinforcement Learning with Expert Guidance
Qisen Yang, Shenzhi Wang, Qihang Zhang +2
Offline reinforcement learning (RL) optimizes the policy on a previously collected dataset without any interactions with the environment, yet usually suffers from the distributiona…