Metraipolis › archive › dc_ffac4298e885
Continual Evolution Strategies in Control Tasks — arXiv:2608.13600
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.13600
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-19T02:30:42.545Z
- SHA-256 of the archived text
9e078666aa45566425fbc857894086752e5a7e2f696a27da278a7f4186697404- Stable URL of this page
https://metraipolis.com/documento/dc_ffac4298e885
How to cite
Continual Evolution Strategies in Control Tasks — arXiv:2608.13600 — arXiv (en), https://arxiv.org/abs/2608.13600 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_ffac4298e885 (Metraipolis archive dc_ffac4298e885, sha256 9e078666aa45566425fbc857894086752e5a7e2f696a27da278a7f4186697404)
BibTeX
@misc{dcffac4298e885,
title = {Continual Evolution Strategies in Control Tasks — arXiv:2608.13600},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.13600},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_ffac4298e885; sha256 9e078666aa45566425fbc857894086752e5a7e2f696a27da278a7f4186697404},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.13600. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
We study Evolution Strategies (ES) for continual control, where agents must adapt to changing tasks without forgetting previous ones. On sequential MuJoCo locomotion tasks, naive ES suffers from severe catastrophic forgetting. Replay substantially improves retention and can induce positive transfer, while larger replay budgets reduce plasticity. Overall, these results show that ES can support continual adaptation in control and that replay is an effective mechanism for mitigating forgetting. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-17 in cs.LG). Categorie arXiv: Neural and Evolutionary Computing (cs.NE) ; Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2608.13600 — 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 2026-10-19T02:26:17.812Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-08-17?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).