Metraipolis › archive › dc_52e2a89566ff

CAST: Alternating State-Value Targets and Expanded Policy Gradients for Model-Based Reinforcement Learning — arXiv:2609.08853

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.08853
License (as archived)
CC0-1.0
Archived
2026-10-16T15:31:26.782Z
SHA-256 of the archived text
a032476c8bfda9ddd9b5d7ec56b8b4ad66490813e06c0172757d478d2986dbeb
Stable URL of this page
https://metraipolis.com/documento/dc_52e2a89566ff

How to cite

CAST: Alternating State-Value Targets and Expanded Policy Gradients for Model-Based Reinforcement Learning — arXiv:2609.08853 — arXiv (en), https://arxiv.org/abs/2609.08853 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_52e2a89566ff (Metraipolis archive dc_52e2a89566ff, sha256 a032476c8bfda9ddd9b5d7ec56b8b4ad66490813e06c0172757d478d2986dbeb)
BibTeX
@misc{dc52e2a89566ff,
  title = {CAST: Alternating State-Value Targets and Expanded Policy Gradients for Model-Based Reinforcement Learning — arXiv:2609.08853},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.08853},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_52e2a89566ff; sha256 a032476c8bfda9ddd9b5d7ec56b8b4ad66490813e06c0172757d478d2986dbeb},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.08853. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Model-based reinforcement learning (MBRL) is a family of RL methods that learn a model of the environment and use it for action selection, making it well suited to robotics due to its sample efficiency. Combining learned models with online planning can further improve action selection, as the planner can exploit the model to find better actions than the learned policy alone. Recent methods combining learned policies with online planning typically learn the value of the policy rather than the stronger planner-guided behavior. We present CAST (Critic with Alternating State-value Target), which uses planner-guided behavior to improve value learning while regularizing the value estimate with the current policy. CAST replaces the action-value critic with a state-value critic, trained using a target that combines a real planner-guided transition and an imagined transition under the current policy. The resulting value function corresponds to an alternating process between planner-guided behavior and the current policy, allowing it to benefit from the stronger planner behavior while being regularised by the policy being learned. We evaluate CAST on the DeepMind Control and HumanoidBench Suites against several state-of-the-art methods, and demonstrate successful transfer to a physical Unitree Go2 quadruped performing a dynamic handstand.

Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-09-09 in cs.RO).
Categorie arXiv: Robotics (cs.RO).
Pagina del lavoro: https://arxiv.org/abs/2609.08853 — 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 2026-10-16T15:11:47.653Z dalla pagina https://arxiv.org/catchup/cs.RO/2026-09-09?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).