Metraipolis › archive › dc_fd2f6fac947b

Beyond Scripted Search: Sample-Efficient Reward Discovery via Agentic Black-box Optimization — arXiv:2609.32394

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.32394
License (as archived)
CC0-1.0
Archived
2026-10-07T20:55:22.719Z
SHA-256 of the archived text
1b3253023afa0974fdff0de2e2c46744c5f029874ecde58c2245f28aeee9c76d
Stable URL of this page
https://metraipolis.com/documento/dc_fd2f6fac947b

How to cite

Beyond Scripted Search: Sample-Efficient Reward Discovery via Agentic Black-box Optimization — arXiv:2609.32394 — arXiv (en), https://arxiv.org/abs/2609.32394 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_fd2f6fac947b (Metraipolis archive dc_fd2f6fac947b, sha256 1b3253023afa0974fdff0de2e2c46744c5f029874ecde58c2245f28aeee9c76d)
BibTeX
@misc{dcfd2f6fac947b,
  title = {Beyond Scripted Search: Sample-Efficient Reward Discovery via Agentic Black-box Optimization — arXiv:2609.32394},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.32394},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_fd2f6fac947b; sha256 1b3253023afa0974fdff0de2e2c46744c5f029874ecde58c2245f28aeee9c76d},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.32394. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Designing dense reward functions for low-level reinforcement learning (RL) control remains difficult. Recent work uses large language models (LLMs) to iteratively generate and refine reward functions using policy-training feedback within scripted search algorithms. However, evaluating each candidate requires a full RL training run, making sample efficiency a central challenge for reward search on complex control tasks. To address this limitation, we propose an Agentic Reward Black-box Optimization (ARBO) framework, in which an LLM agent builds the search strategy at run time from an evaluation history maintained as its persistent workspace. The evaluation history comprises two components: observations maintained by the evaluation oracle, including candidate scores, per-term training curves, and error tracebacks; and an agent-maintained belief that records diagnoses and intended next steps. The agent queries both with tools and generates the next batch of reward candidates, rather than generating them in a single pass from a fixed prompt. Across four control domains, ARBO achieves gains of 29.9% in manipulation success rate and 192.8% in power-grid score over baseline means under a shared evaluation budget. Ablations examine each component's contribution and sensitivity to backbone choice.

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/2609.32394 — 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 2026-10-07T20:29:00.777Z 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).