Metraipolis › archive › dc_2dfda995a0ad
Runtime-Incremental Transformer for Reinforcement-Learning-Based Adaptive Control — arXiv:2609.13545
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.13545
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-16T00:12:48.785Z
- SHA-256 of the archived text
cee8b59256ca9eafd2305aeef2324b5412fc79db495a8975c31c5740b85df5e4- Stable URL of this page
https://metraipolis.com/documento/dc_2dfda995a0ad
How to cite
Runtime-Incremental Transformer for Reinforcement-Learning-Based Adaptive Control — arXiv:2609.13545 — arXiv (en), https://arxiv.org/abs/2609.13545 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_2dfda995a0ad (Metraipolis archive dc_2dfda995a0ad, sha256 cee8b59256ca9eafd2305aeef2324b5412fc79db495a8975c31c5740b85df5e4)
BibTeX
@misc{dc2dfda995a0ad,
title = {Runtime-Incremental Transformer for Reinforcement-Learning-Based Adaptive Control — arXiv:2609.13545},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.13545},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_2dfda995a0ad; sha256 cee8b59256ca9eafd2305aeef2324b5412fc79db495a8975c31c5740b85df5e4},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.13545. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Learning-based adaptive control of robotic manipulators with non-observable friction memory has been addressed by attention- based meta-controllers whose number of attention heads is fixed before training and is tuned by costly offline search. At long memory horizons, such fixed-capacity controllers are prone to catastrophic failures on a sizeable fraction of training seeds. The present paper introduces a runtime mechanism that grows and prunes the heads of the attention block during reinforcement learning, governed by two signals: the effective rank of the on-policy context distribution, which triggers growth when representational capacity becomes insufficient, and the per-head output magnitude, which flags redundant heads for removal. Policy continuity at growth events and a quantitative bound at prune events are established analytically. On a two- link manipulator with Stribeck friction, the proposed mechanism attains full success across all memory regimes, eliminating the long-horizon failure mode and removing the need for offline tuning of the head count. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-15 in cs.LG). Categorie arXiv: Robotics (cs.RO) ; Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2609.13545 — 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 2026-10-15T23:54:14.849Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-09-15?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).