Newest papers

EventSync: Event-Driven Memory Updates for History-Dependent Manipulation

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

Indonesian

Manipulasi robot yang bergantung pada sejarah memerlukan konteks tugas dan sejarah interaksi, karena pengamatan saat ini saja tidak cukup untuk menentukan tindakan berikutnya. Kerangka kerja hierarki menggunakan memori tugas untuk alasan subtarget tingkat tinggi dan kebijakan tingkat rendah untuk eksekusi. Namun, pembaruan memori membutuhkan penilaian penutupan yang akurat: pengamatan serupa mungkin sesuai dengan fase eksekusi yang berbeda, dan pencatatan prematur dapat menyesatkan penilaian subtarget berikutnya. Kami menyajikan EventSync, kerangka kerja hierarki yang mengkoordinasikan penilaian subtarget dan eksekusi tindakan melalui pembaruan memori yang didorong oleh peristiwa. Model vision-language membutuhkan penilaian penutupan yang akurat: pengamatan serupa dapat sesuai dengan fase eksekusi yang berbeda, sementara model vision-language-action mengeksekusi mereka. Sebuah Monitor Event ringan menggabung

Machine translation (Meta M2M100); the English original follows.

Abstract

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)

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange