Metraipolis › archive › dc_0082aa160edf

Externalise the Memory: A Pre-Registered Study of Where an Embodied Agent's Experience Should Live — arXiv:2610.30478

Source: arXiv (en)
Original: https://arxiv.org/abs/2610.30478
License (as archived): CC0-1.0
Archived: 2026-10-23T15:14:56.714Z
SHA-256 of the archived text: e131bc8ce7da21f93f75aaf17bb418badaa0369cb213aad58300911382da9996
Stable URL of this page: https://metraipolis.com/documento/dc_0082aa160edf

How to cite

Externalise the Memory: A Pre-Registered Study of Where an Embodied Agent's Experience Should Live — arXiv:2610.30478 — arXiv (en), https://arxiv.org/abs/2610.30478 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_0082aa160edf (Metraipolis archive dc_0082aa160edf, sha256 e131bc8ce7da21f93f75aaf17bb418badaa0369cb213aad58300911382da9996)

BibTeX

@misc{dc0082aa160edf,
  title = {Externalise the Memory: A Pre-Registered Study of Where an Embodied Agent's Experience Should Live — arXiv:2610.30478},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.30478},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_0082aa160edf; sha256 e131bc8ce7da21f93f75aaf17bb418badaa0369cb213aad58300911382da9996},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.30478. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Should an embodied agent's experience be baked into its policy weights, or kept external and retrieved at inference? A memory-augmented teacher (a VLM that retrieves lessons, with their episode images, from its own past attempts) produces trajectories on a vision-conditioned reach task; from the same trajectories we distil LoRA student policies and evaluate them under matched protocols, including one that re-attaches the teacher's frozen retrieval buffer to a fixed set of student weights. We pre-registered the hypotheses, arms, tests, and analysis rules before training and before any p-value. Distilling the memory-augmented behaviour into weights beat an action-only control by +1 pp (n.s.); the registered test proves underpowered for its own minimum effect, so we withdraw - rather than refute - the memory-in-weights claim. Attaching retrieval to a fixed set of student weights added +34 pp (z=5.15), reaching the pooled memory-augmented teacher reference (49.3%); the pre-registered protocol contrast agrees (+23 pp, z=3.36); under a pre-registered auto-reset sensitivity count the two are +36 and +29 pp. An exploratory behavioural probe with pre-registered per-arm predictions is consistent with behavioural distillation transferring the marginal correction but not the conditional, situation-specific correction that retrieval supplies. The recipe that follows is to externalise the memory: in follow-up controls registered before they ran, the same frozen buffer lifts the base model with no adapter to 48% (sensitivity 48%) and the other three adapters from 19-26% to 53-55% (sensitivity 17-21% to 50-55%), and the adapter adds nothing detectable once the buffer is attached. On a second, harder task, the retrieval advantage points the same direction but falls short of significance (+6 pp, 95% CI [-4.5, +16.4]); exploratory diagnostics tie the attenuation to memory coverage.

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/2610.30478 — 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 2026-10-23T15:09:08.524Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).