8 papers
Chunk-wise Model-based Policy Optimization
Weixi Song, Donglin Wang, Yong-Lu Li
Model-based reinforcement learning has shown superior sample efficiency by leveraging the learned latent dynamics either in the form of derivative-free planning/imagined rollout tr…
Explicit Kinematic Guidance from Analytic Concepts for Vision-Language-Action Models
Mingyang Sun, Jiude Wei, Xiujian Liang +4
Current Vision-Language-Action (VLA) models rely mainly on 2D inputs, neglecting the rich object structural information and commonsense knowledge inherent in the 3D physical world.…
Analytic Concept-Centric Memory for Agentic Embodied Manipulation
Mingyang Sun, Xiujian Liang, Jiude Wei +4
Long-horizon embodied manipulation requires agents to remember persistent objects, track changing scene states, and reuse prior interaction knowledge. However, existing agent memor…
Rethinking the Practicality of Vision-language-action Model: A Comprehensive Benchmark and An Improved Baseline
Wenxuan Song, Jiayi Chen, Xiaoquan Sun +12
Vision-Language-Action (VLA) models have emerged as a generalist robotic agent. However, existing VLAs are hindered by excessive parameter scales, prohibitive pre-training requirem…
L1 Sample Flow for Efficient Visuomotor Learning
Weixi Song, Zhetao Chen, Tao Xu +6
Denoising-based models, such as diffusion and flow matching, have been a critical component of robotic manipulation for their strong distribution-fitting and scaling capacity. Conc…
Iterative Refinement of Flow Policies in Probability Space for Online Reinforcement Learning
Mingyang Sun, Pengxiang Ding, Weinan Zhang +1
While behavior cloning with flow/diffusion policies excels at learning complex skills from demonstrations, it remains vulnerable to distributional shift, and standard RL methods st…
Executable Analytic Concepts as the Missing Link Between VLM Insight and Precise Manipulation
Mingyang Sun, Jiude Wei, Qichen He +3
Enabling robots to perform precise and generalized manipulation in unstructured environments remains a fundamental challenge in embodied AI. While Vision-Language Models (VLMs) hav…
TrajBooster: Boosting Humanoid Whole-Body Manipulation via Trajectory-Centric Learning
Jiacheng Liu, Pengxiang Ding, Qihang Zhou +8
Recent Vision-Language-Action models show potential to generalize across embodiments but struggle to quickly align with a new robot's action space when high-quality demonstrations…