Metraipolis › archive › dc_bb8120207c8b

PhysEvo: Astra Can Act, Let It — arXiv:2610.08995

Source: arXiv (en)
Original: https://arxiv.org/abs/2610.08995
License (as archived): CC0-1.0
Archived: 2026-10-08T11:10:27.274Z
SHA-256 of the archived text: e534b44790b4909d20827f85d50fdb0b42c19edba55b82d40d043851d23dfb8f
Stable URL of this page: https://metraipolis.com/documento/dc_bb8120207c8b

How to cite

PhysEvo: Astra Can Act, Let It — arXiv:2610.08995 — arXiv (en), https://arxiv.org/abs/2610.08995 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_bb8120207c8b (Metraipolis archive dc_bb8120207c8b, sha256 e534b44790b4909d20827f85d50fdb0b42c19edba55b82d40d043851d23dfb8f)

BibTeX

@misc{dcbb8120207c8b,
  title = {PhysEvo: Astra Can Act, Let It — arXiv:2610.08995},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.08995},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_bb8120207c8b; sha256 e534b44790b4909d20827f85d50fdb0b42c19edba55b82d40d043851d23dfb8f},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.08995. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

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 manipulation skills without model-weight updates or a separately trained action policy. Across 42 RoboDojo tasks, held-out-layout evaluation of retained task-specific deployment versions yields a five-dimension average score of 68.14/100 and 62.00% success, compared with 47.17% for RoboDawn's one-shot Astra agent, the strongest published reference in our comparison. On eight manipulation tasks challenging direct Astra, PhysEvo achieves 55.00% success, compared with 1.25% for the direct-Astra reference. Deploying the simulation-evolved harness on AgileX PiPER and continuing skill revision yields 90.60/100 average score and 84.00% success across 25 trials on five real-world tasks. PhysEvo turns the consequences of action into persistent, testable changes to how a frozen model acts and improves.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Robotics (cs.RO) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2610.08995 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-08T10:14:06.037Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).