Bofang Jia

Zeyi Liu, Guangyao Liu, Yinuo Qu et al.

Human-in-the-loop reinforcement learning (HIL-RL) improves sample efficiency in real-robot manipulation through online human intervention. However, successful trajectories may include suboptimal actions that deviate from the desired task-execution path and force human intervention. Existing HIL-RL methods typically apply the consistent credit assignment principle to all transitions, uniformly propagating discounted terminal rewards through suboptimal segments, ignoring the actual contribution of each transition to task success. This overestimates Q-values for critic learning and indirectly misguides actor updates toward suboptimal behavior patterns. To this end, we propose PACT, a Preference-calibrated Actor-Critic Training framework that leverages the implicit preference signals induced by intervention to perform credit reassignment on identified suboptimal segments while directly guiding policy training for unbiased critic-actor learning. Specifically, we first design a progress model that learns from human demonstration and identifies suboptimal segments for credit correction. Then, from the human action and resampled policy action at the intervention state, we build preference pairs to define a counterfactual advantage that penalizes Bellman targets of the identified suboptimal segment, enabling directional credit calibration. Moreover, we directly align the policy with human corrective actions in the bounded mean space, providing an additional signal beyond critic-guided updates. Across five real-robot manipulation tasks, PACT improves the average success rate by 24.5% and achieves 1.3 times faster convergence, thereby improving both RL sample efficiency and performance. Code is available at https://anonymous.4open.science/r/HILRL-A1X-BC05.

Can Cui, Pengxiang Ding, Wenxuan Song et al.

Dual-system VLA (Vision-Language-Action) architectures have become a hot topic in embodied intelligence research, but there is a lack of sufficient open-source work for further performance analysis and optimization. To address this problem, this paper will summarize and compare the structural designs of existing dual-system architectures, and conduct systematic empirical evaluations on the core design elements of existing dual-system architectures. Ultimately, it will provide a low-cost open-source model for further exploration. Of course, this project will continue to update with more experimental conclusions and open-source models with improved performance for everyone to choose from. Project page: https://openhelix-robot.github.io/.

Bofang Jia, Pengxiang Ding, Can Cui et al.

Visual-motor policy learning has advanced with architectures like diffusion-based policies, known for modeling complex robotic trajectories. However, their prolonged inference times hinder high-frequency control tasks requiring real-time feedback. While consistency distillation (CD) accelerates inference, it introduces errors that compromise action quality. To address these limitations, we propose the Score and Distribution Matching Policy (SDM Policy), which transforms diffusion-based policies into single-step generators through a two-stage optimization process: score matching ensures alignment with true action distributions, and distribution matching minimizes KL divergence for consistency. A dual-teacher mechanism integrates a frozen teacher for stability and an unfrozen teacher for adversarial training, enhancing robustness and alignment with target distributions. Evaluated on a 57-task simulation benchmark, SDM Policy achieves a 6x inference speedup while having state-of-the-art action quality, providing an efficient and reliable framework for high-frequency robotic tasks.

Yuquan Xue, Guanxing Lu, Zhenyu Wu et al.

Vision-Language-Action models (VLAs) have demonstrated remarkable performance on complex robotic manipulation tasks through imitation learning. However, existing imitation learning datasets contain only successful trajectories and lack failure or recovery data, especially for out-of-distribution (OOD) states where the robot deviates from the main policy due to minor perturbations or errors, leading VLA models to struggle with states deviating from the training distribution. To this end, we propose an automated OOD data augmentation framework named RESample through exploratory sampling. Specifically, we first leverage offline reinforcement learning to obtain an action-value network that accurately identifies sub-optimal actions under the current manipulation policy. We further sample potential OOD states from trajectories via rollout, and design an exploratory sampling mechanism that adaptively incorporates these action proxies into the training dataset to ensure efficiency. Subsequently, our framework explicitly encourages the VLAs to recover from OOD states and enhances their robustness against distributional shifts. We conduct extensive experiments on the LIBERO benchmark as well as real-world robotic manipulation tasks, demonstrating that RESample consistently improves the stability and generalization ability of VLA models.