Metraipolis › archive › dc_536d4ed7ffc3

Density-Aware Reward Scaling: A Reinforcement Learning Framework for Reward-Proportional Sampling — arXiv:2611.04792

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.04792
License (as archived)
CC0-1.0
Archived
2026-11-04T12:30:36.892Z
SHA-256 of the archived text
1d365f492a14b0252c49dce9b1f18a9ca9c050b357ef9fca4f9d558e496ecac7
Stable URL of this page
https://metraipolis.com/documento/dc_536d4ed7ffc3

How to cite

Density-Aware Reward Scaling: A Reinforcement Learning Framework for Reward-Proportional Sampling — arXiv:2611.04792 — arXiv (en), https://arxiv.org/abs/2611.04792 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_536d4ed7ffc3 (Metraipolis archive dc_536d4ed7ffc3, sha256 1d365f492a14b0252c49dce9b1f18a9ca9c050b357ef9fca4f9d558e496ecac7)
BibTeX
@misc{dc536d4ed7ffc3,
  title = {Density-Aware Reward Scaling: A Reinforcement Learning Framework for Reward-Proportional Sampling — arXiv:2611.04792},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.04792},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_536d4ed7ffc3; sha256 1d365f492a14b0252c49dce9b1f18a9ca9c050b357ef9fca4f9d558e496ecac7},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.04792. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Reinforcement learning objectives that maximize expected return can concentrate probability mass on a small subset of high-reward outcomes, limiting diversity when multiple distinct solutions are desirable. We introduce Density-Aware Reward Scaling (DARS), a simple reward-processing mechanism that promotes reward-proportional sampling in continuous outcome spaces while retaining standard policy-gradient training. DARS estimates the local density of outcomes within each on-policy rollout group and inversely scales rewards by this density. The method requires no learned density model, auxiliary objective, critic, or modification to the underlying policy-gradient objective. In a controlled multimodal navigation task, DARS-GRPO avoids the single-mode collapse exhibited by GRPO and closely recovers the target mode allocation. Applied to text-to-image prompt adaptation, DARS-GRPO consistently improves image and prompt diversity across datasets, Policy LMs, and reward settings while maintaining good semantic alignment. These results show that density-aware reward scaling provides a lightweight way to improve coverage of diverse high-reward outcomes within existing reinforcement-learning pipelines.

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/2611.04792 — 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 2026-11-04T12:26:47.297Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).