Metraipolis › archive › dc_5712375348ee
Evolving Support Priorities in Empathetic Reinforcement Learning — arXiv:2609.34249
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.34249
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-07T20:57:34.228Z
- SHA-256 of the archived text
2155453ace724e0ba8a7d9204089ab32ab5fe512740e01051a04e7590209be04- Stable URL of this page
https://metraipolis.com/documento/dc_5712375348ee
How to cite
Evolving Support Priorities in Empathetic Reinforcement Learning — arXiv:2609.34249 — arXiv (en), https://arxiv.org/abs/2609.34249 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_5712375348ee (Metraipolis archive dc_5712375348ee, sha256 2155453ace724e0ba8a7d9204089ab32ab5fe512740e01051a04e7590209be04)
BibTeX
@misc{dc5712375348ee,
title = {Evolving Support Priorities in Empathetic Reinforcement Learning — arXiv:2609.34249},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.34249},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_5712375348ee; sha256 2155453ace724e0ba8a7d9204089ab32ab5fe512740e01051a04e7590209be04},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.34249. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
We identify a fundamental mismatch in empathetic reinforcement learning: support priorities evolve with the dialogue state, yet existing methods typically optimize predefined reward specifications that remain fixed across turns. To model these evolving support priorities, we organize empathetic support along cognitive, affective, and proactive empathy, and propose Context-Adaptive Rubric Evolution (CARE). At each turn, CARE generates a context-adaptive rubric by adjusting both the weights of these three empathy dimensions and their fine-grained evaluation criteria. The rubric generator is trained with turn-level rubric supervision and human preference data through supervised fine-tuning followed by preference-based reinforcement learning, and then serves as an adaptive reward interface for online empathetic RL. Integrated with both RLVER and MICA, CARE achieves state-of-the-art performance across SentientBench, EQBench3, and EMPA under three independent LLM judges. Notably, on EMPA, CARE improves EPM-Idx over the strongest baseline by at least 13 points under all three judges, including an increase from 28.11 to 83.54 under Gemini-2.5-Pro. Further analyses show that learned rubric priorities systematically vary across dialogue stages and user emotions, demonstrating that CARE adapts what is rewarded as support needs evolve. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-29 in cs.AI). Categorie arXiv: Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2609.34249 — 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 2026-10-07T20:29:01.491Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).