PhysEvo: Astra Can Act, Let It

类型：arxiv
标识：2610.08995
链接：https://arxiv.org/abs/2610.08995
主分类：agent
形态：method
TLDR：Astra can act, yet reliable manipulation depends on the system through which it observes and controls the world. We introduce PhysEvo, a framework for physical recursive self-improvement (RSI) around a single frozen model. A task agent executes robot tasks; a meta-agent uses the resulting trajectories to diagnose failures, revise tools and skills, and test corrections. The meta-agent can also improve its own diagnostic tools, so retained revisions support both later action and later self-improvement. This process develops joint-level control, evidence-seeking observation, and reusable manipula
待LLM分类：否
来源文件：
/inbox/tom/_candidates/2026-10-08-agent-rag-longcontext-candidates.json

2026-08-10

笔记 Microsoft Research Blog · RSS 摘要

2026-08-12

笔记 AgentRx 精读与批判（v2 · 覆盖重写版）— flyP

2026-07-25

笔记精读与批判：AgentRx —— 多模态临床预测中，单 Agent 反超多 Agent

2026-06-30

笔记 2026-07-04 上午轻量精读 · AgenticRAGTracer（hop-aware 多跳诊断基准）

2026-07-04

笔记 Cool Papers · cs.CL (papers.cool) · RSS 摘要

2026-08-20

arXiv

2610.08995

TLDR

Astra can act, yet reliable manipulation depends on the system through which it observes and controls the world. We introduce PhysEvo, a framework for physical recursive self-improvement (RSI) around a single frozen model. A task agent executes robot tasks; a meta-agent uses the resulting trajectories to diagnose failures, revise tools and skills, and test corrections. The meta-agent can also improve its own diagnostic tools, so retained revisions support both later action and later self-improvement. This process develops joint-level control, evidence-seeking observation, and reusable manipula

PhysEvo: Astra Can Act, Let It

相关 · 按语义匹配