Metraipolis › archive › dc_0aa5867ea57d

METTA: Memory-Enhanced Test-Time Adaptation for Meta-Reinforcement Learning — arXiv:2610.31656

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.31656
License (as archived)
CC0-1.0
Archived
2026-10-23T15:16:49.574Z
SHA-256 of the archived text
a41868e2de32d19cd6249aeea933a0095d9dd46049f8ca06bd8079f264987992
Stable URL of this page
https://metraipolis.com/documento/dc_0aa5867ea57d

How to cite

METTA: Memory-Enhanced Test-Time Adaptation for Meta-Reinforcement Learning — arXiv:2610.31656 — arXiv (en), https://arxiv.org/abs/2610.31656 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_0aa5867ea57d (Metraipolis archive dc_0aa5867ea57d, sha256 a41868e2de32d19cd6249aeea933a0095d9dd46049f8ca06bd8079f264987992)
BibTeX
@misc{dc0aa5867ea57d,
  title = {METTA: Memory-Enhanced Test-Time Adaptation for Meta-Reinforcement Learning — arXiv:2610.31656},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.31656},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_0aa5867ea57d; sha256 a41868e2de32d19cd6249aeea933a0095d9dd46049f8ca06bd8079f264987992},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.31656. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Meta-reinforcement learning aims to train agents that adapt quickly to unseen tasks from limited interaction, but this requires representing task evidence over long interaction histories. Classical recurrent agents compress experience into hidden states, while transformer-based agents retrieve from past trajectories at quadratic memory and computation cost. In this work, we propose Memory-Enhanced Test-Time Adaptation (METTA) for meta-reinforcement learning. It decomposes the fast adaptation phase into two temporal scales: a causal neural memory performs fine-grained within-episode adaptation with linear sequence complexity, while compact episode-level summaries support long-range cross-episode task inference without attending over the full interaction history. Moreover, equipping agents with neural memory provides sufficient expressive capacity and test-time training capability, enabling effective adaptation to unseen held-out tasks and task families at meta-test time. On unseen held-out Meta-World ML10/45 test environments, METTA achieves a success rate over $50\%$ while maintaining a low training budget and short wall-clock time. On long-context sparse-reward PointMaze tasks, it outperforms transformer-based baselines by exhibiting more structured exploration behavior. These results show that METTA provides an efficient path toward more adaptive and scalable long-context meta-RL.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2610.31656 — 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 2026-10-23T15:09:08.646Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).