Pro$^2$ Agent is introduced, which shifts the primary state of autonomous AI research from a codebase state to a Problem State: an explicit structure of what problems prevent success and which solutions are likely to address them.
Yi-Xiong Chen, Li-Qian Peng, Tian Tan et al.· 0 citations
Enter--Propagation--Recovery (E-P-R), a trajectory-level framework that asks where memory first changes an action, whether that change carries forward, and whether the agent can recover after leaving a correct path is proposed, to diagnose how models consume retrieved memory across a multi-step action trajectory.
Yi-Xiong Chen, Xin-Yi Bai, Alan L. Yuille· arXiv.org· 0 citations
Agentic Real2Sim is introduced, a framework for generalized physical world modeling with vision-language agents, converting a real-world recording of object-robot interaction into a simulatable episodic twin which preserves observations, geometries, robot interactions, and object states.