12 papers
PRISM: Precision Forcing for Generalist Vision-Language-Action Policies
Chenghao Yue, Peiyan Li, Hao Wang +5
Coupling a pretrained vision-language model (VLM) with a diffusion or flow-matching action expert has become the mainstream vision-language-action (VLA) paradigm for embodied polic…
TimelyDAgger: Timing-Aware Expert Querying for VLA Policy Improvement
Zhixuan Zhao, Peiyan Li, Enhao Zhang +10
DAgger improves robot policies by aggregating expert supervision from states visited during policy execution. Robot-gated DAgger automates expert queries, allowing the robot to dec…
Beyond State-as-Action: Exploiting Command-State Discrepancy for Robot Imitation Learning
Peiyan Li, Yueran Tao, Enhao Zhang +10
Constructing action targets from measured robot motion is an established approach in imitation learning. Under interaction constraints, however, command-state discrepancy may refle…
Social World Model for Lifelong Social Intelligence
Yu Luo
Social intelligence is a core competency for language agents, yet current research primarily focuses on static capability evaluation rather than how these skills are continuously s…
FLAC: Maximum Entropy RL via Kinetic Energy Regularized Bridge Matching
Lei Lv, Yunfei Li, Yu Luo +2
Iterative generative policies, such as diffusion models and flow matching, offer superior expressivity for continuous control but complicate Maximum Entropy Reinforcement Learning…
Flow-Based Policy for Online Reinforcement Learning
Lei Lv, Yunfei Li, Yu Luo +4
We present \textbf{FlowRL}, a novel framework for online reinforcement learning that integrates flow-based policy representation with Wasserstein-2-regularized optimization. We arg…
Multi-segment Soft Robot Control via Deep Koopman-based Model Predictive Control
Lei Lv, Lei Liu, Lei Bao +7
Soft robots, compared to regular rigid robots, as their multiple segments with soft materials bring flexibility and compliance, have the advantages of safe interaction and dexterou…
Towards Task-Oriented Flying: Framework, Infrastructure, and Principles
Kangyao Huang, Hao Wang, Jingyu Chen +6
Deploying robot learning methods to aerial robots in unstructured environments remains both challenging and promising. While recent advances in deep reinforcement learning (DRL) ha…
Bidirectional-Reachable Hierarchical Reinforcement Learning with Mutually Responsive Policies
Yu Luo, Fuchun Sun, Tianying Ji +1
Hierarchical reinforcement learning (HRL) addresses complex long-horizon tasks by skillfully decomposing them into subgoals. Therefore, the effectiveness of HRL is greatly influenc…
OMPO: A Unified Framework for RL under Policy and Dynamics Shifts
Yu Luo, Tianying Ji, Fuchun Sun +3
Training reinforcement learning policies using environment interaction data collected from varying policies or dynamics presents a fundamental challenge. Existing works often overl…
Offline-Boosted Actor-Critic: Adaptively Blending Optimal Historical Behaviors in Deep Off-Policy RL
Yu Luo, Tianying Ji, Fuchun Sun +3
Off-policy reinforcement learning (RL) has achieved notable success in tackling many complex real-world tasks, by leveraging previously collected data for policy learning. However,…
Smooth Computation without Input Delay: Robust Tube-Based Model Predictive Control for Robot Manipulator Planning
Yu Luo, Qie Sima, Tianying Ji +3
Model Predictive Control (MPC) has exhibited remarkable capabilities in optimizing objectives and meeting constraints. However, the substantial computational burden associated with…