This work studies HIRO-style hierarchy, in which a high-level policy proposes subgoals for a goal-conditioned low-level policy and an off-policy correction relabels past subgoals as the worker improves, and studies an object-centric variant, in which subgoals are defined as relative vectors between task-relevant entities rather than as raw, embodiment-specific robot states.
This thesis investigates what the high level goal predictor should predict, how the low-level policy should consume it, what capabilities it must encode, and how a user might steer it all remain open through three studies that share one object of study: the high-level goal predictor.
This study proposes a new method, Wrench-Augmented Reinforcement Learning (WARL), which introduces a wrenche (force and torque) into the action space, and shows that introducing a wrench can encourage behaviors that do not sufficiently exploit the robot's physical embodiment.
Two systematic attempts to improve large pretrained models with minimal or zero modification to their weights via reinforcement learning on a frozen OpenVLA-7B using binary task-success rewards on LIBERO-Goal reveal a common ceiling.
The results provide stronger evidence that RL-only bootstrapping can create the first usable language-conditioned controller for a genuinely novel embodiment, and use reinforcement learning in simulation with dense geometric rewards computed from simulator state.
Damir Nurtdinov, A. Kornaev, A. Maloletov· 0 citations
A unified framework that combines centralized training with decentralized execution (CTDE) and a Hybrid Reward Architecture (HRA) is introduced that enables multiple actors to share a centralized multi-head critic and substantially improves both sample efficiency and policy performance.
Changhao Li, Yifang Zhang, Heng Zhang et al.· 0 citations
Learning long-horizon robot manipulation remains difficult and time-consuming, especially under sparse rewards due to inefficient exploration and reward assignment. We present a minimal integration of large language models (LLMs) with reinforcement learning (RL) in which the LLM is used strictly as an online action pro...
Meiyuan Gong, Yan Gao, Ze Ji· 2026 IEEE International Conf...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
