Jingjiao Zhao, Jing Dong, Mingyu Cao, Ruochuan Shi, Xiansheng Chen, Pengwei Wang, Zhiwu Lu · 2026 · arXiv (Cornell University)
DOI: 10.48550/arxiv.2610.27917 · cited by 0 works (OpenAlex) · open-access PDF
La manipulation robotique dépendante de l'histoire nécessite un contexte de tâches et un historique d'interaction, car l'observation actuelle seule n'est pas suffisante pour déterminer l'action suivante. Les cadres hiérarchiques utilisent la mémoire de tâche pour le raisonnement de sous-objectifs de haut niveau et les politiques de bas niveau pour l'exécution. Cependant, les mises à jour de mémoire nécessitent une évaluation précise de l'achèvement: des observations similaires peuvent correspondre à différentes phases d'exécution, tandis qu'un modèle d'action d'enregistrement prématuré peut induire en erreur le raisonnement de sous-objectif ultérieur. Nous présentons EventSync, un cadre hiérarchique
Machine translation (Meta M2M100); the English original follows.
History-dependent robotic manipulation requires task context and interaction history, as the current observation alone is insufficient to determine the next action. Hierarchical frameworks use task memory for high-level subgoal reasoning and low-level policies for execution. However, memory updates require accurate completion assessment: similar observations may correspond to different execution phases, and prematurely recording completion can mislead subsequent subgoal reasoning. We present EventSync, a hierarchical framework that coordinates task reasoning and action execution through event-driven memory updates. A vision-language model generates subgoals from task plans and interaction history, while a vision-language-action model executes them. A lightweight Event Monitor combines visual observations with end-effector trajectories to determine whether the current subgoal is complete. The framework records the current subgoal in its completed-subgoal history and triggers subsequent reasoning only after obtaining completion evidence. With visual history in the executor, EventSync achieves 64.5% overall success on RoboMME, exceeding the best baseline reported in the benchmark by 20.0 percentage points. Further analysis shows that this update mechanism helps build a reliable completed-subgoal history. Experiments on four real-robot tasks with different memory requirements validate the framework's effectiveness in real-world robotic manipulation.
Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese
Topics: Computer science, Task (project management), Context (archaeology), Artificial intelligence, Benchmark (surveying), Action (physics)
GET https://store.agentexchange.work/gw/papers/verify?title=...&year=... (0.003 USD), fetch this record as JSON with /gw/papers/paper?id=2610.27917, or list what it cites with /gw/papers/references?id=.... Pay per call in USDC (x402) or with a 19 USD card key for 1,000 calls. Catalogue · llms.txtSources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange