AIAny
Icon for item

Envisioning Agentic Robotics: A New Era of Robotic Systems

Proposes a framework for robots that combine foundation models with an embodied runtime to interpret intent, call skills, verify outcomes, and adapt. It also outlines an evaluation framework and a roadmap toward sustained autonomy.

Introduction

A robot that can choose actions is not necessarily an autonomous agent: it must also determine whether those actions worked and decide what to do next. This paper makes that closed-loop responsibility—not the use of a particular model or planner—the defining test for agentic robotics, then uses it to organize evaluation and research priorities.

Information

  • Websitearxiv.org
  • OrganizationsGrid Robotics
  • AuthorsQingfeng Lan, Duo Li, Siyao Li, Yiheng Lyu, Shijian Lu, Zhenghao Qi, Shizun Wang, Xinchao Wang, Ming Yang, Zhiqin Yang …
  • Published date2026/11/13

More Items

Reduces Transformer parameter count by repeatedly applying shared blocks, then combines their updates with geometric decay and learned attention. On ClimbMix, it beats standard baselines in validation loss with fewer Transformer parameters.

Turns screen observations into keyboard and mouse actions to complete long-horizon game tasks. Sparse reasoning, persistent memory, and online reinforcement learning support real-time interaction and multi-step decisions.

Tests decision models as actors, trajectory judges, and reinforcement-learning teachers for long-horizon agents. It finds that a model can be a weak actor yet still improve policy learning by evaluating actions after seeing their consequences.