27 January 2026 in IEEE Transactions on Robotics
Learning from demonstration is a powerful method for robotic skill acquisition. Nevertheless, a critical limitation lies in the substantial costs associated with gathering demonstration datasets, typically action-labeled robot data, which creates a fundamental constraint in the field. Video data offer a compelling solution as an alternative rich data source, containing diverse behavioral and physical knowledge. This study introduces G3M, an innovative framework that exploits video data via Graph-to-Graphs Generative Modeling, which pre-trains models to generate future graphs conditioned on the graph within a video frame. The proposed G3M abstracts video frame into graph representations by identifying object and visual action vertices for capturing state information. It then effectively models internal structures and spatial relationships present in these graph constructions, with the objective of predicting forthcoming graphs. The generated graphs function as conditional inputs that guide the control policy in determining robotic behaviors. This concise method effectively encodes critical spatial relationships while facilitating accurate prediction of subsequent graph sequences, thus allowing the development of resilient control policy despite constraints in action-annotated training samples. Furthermore, these transferable graph representations enable the effective extraction of manipulation knowledge through human videos as well as recordings from robots with different embodiments. The experimental results demonstrate that G3M attains superior performance using merely 20% action-labeled data relative to comparable approaches. Moreover, our method outperforms the state-of-the-art method, showing performance gains exceeding 19% in simulated environments and 23% in real-world experiments, while delivering improvements of over 35% in cross-embodiment transfer experiments and exhibiting strong performance on long-horizon tasks. Our project page is available at https://g3m-project.github.io/.
1 January 2026 in IEEE Transactions on Automation Science and Engineering
9 October 2025 in IEEE Robotics and Automation Letters
While autoregressive models demonstrate remarkable success in text and image generation, their application to robot policies suffers from weak holistic comprehension, cumulative errors, and limited multi-modal modeling capabilities, particularly in long-horizon tasks or multi-modal scenarios. This paper presents HAMR,, a novel Hierarchical Autoregressive framework with Multi-scale action Refinement that addresses these limitations through three key contributions: (1) a hierarchical architecture wherein high-level long-horizon motion features guide low-level fine-grained action generation, (2) a multi-scale action tokenization and generation mechanism that actively corrects previous prediction errors during inference, and (3) a lightweight diffusion-based decoder that enhances multi-modal modeling without compromising efficiency. HAMR achieves 12% and 15% improvements over state-of-the-art methods across 60+ tasks in simulation and real-world experiments, respectively, with a 25% improvement in perturbed experiments, demonstrating superior performance in both short- and long-horizon manipulation tasks.
1 June 2026 in IEEE Robotics and Automation Letters
Real-world reinforcement learning enables robots to adapt directly to complex physical dynamics, avoiding the reality gap inherent in simulation and rendering previously infeasible manipulation skills a reality. However, the high cost of real-world exploration strictly limits sample availability. Such tight interaction budgets make sparse rewards insufficient for sample-efficient learning in real-robot arm manipulation, creating the need for dense reward supervision. Although pretrained reward models provide useful semantic priors, their generic predictions are often misaligned with embodiment-specific and online dynamics during physical deployment. To address this issue, we propose adaptive reward via human interaction (ARHI), a parameter-efficient framework for online adaptation of VLM-based rewards in real-robot reinforcement learning. Rather than treating the pretrained reward as a fixed module, ARHI continuously recalibrates it using lightweight online updates, while preserving the semantic priors of the frozen backbone. To support this online reward adaptation under sparse human feedback, we further integrate the method into an asynchronous dual-loop system that converts sparse interventions into dense supervisory signals for reward and policy learning. Extensive real-world experiments across six manipulation tasks demonstrate that ARHI outperforms strong baselines, reducing human effort by 28.0%, while achieving a 47.6% reduction in performance drop under challenging deployment shifts.