Metraipolis › archive › dc_67f0ec299bc5

Test-Time Scaling for Scientific Equation Discovery — arXiv:2608.28660

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.28660
License (as archived)
CC0-1.0
Archived
2026-10-17T07:47:13.318Z
SHA-256 of the archived text
2e8f75d7a147e38b76f2099827abef408b21c1b21b31fd5cc96b09abb85f51b4
Stable URL of this page
https://metraipolis.com/documento/dc_67f0ec299bc5

How to cite

Test-Time Scaling for Scientific Equation Discovery — arXiv:2608.28660 — arXiv (en), https://arxiv.org/abs/2608.28660 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_67f0ec299bc5 (Metraipolis archive dc_67f0ec299bc5, sha256 2e8f75d7a147e38b76f2099827abef408b21c1b21b31fd5cc96b09abb85f51b4)
BibTeX
@misc{dc67f0ec299bc5,
  title = {Test-Time Scaling for Scientific Equation Discovery — arXiv:2608.28660},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.28660},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_67f0ec299bc5; sha256 2e8f75d7a147e38b76f2099827abef408b21c1b21b31fd5cc96b09abb85f51b4},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.28660. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Test-time scaling (TTS) improves language model reasoning by allocating additional test-time compute, but prior work mainly studies closed-ended tasks such as math and coding. We study TTS for automated equation discovery, an open-ended setting where models search over candidate equations and rely on observed datapoints for feedback. We formulate LLM-driven equation discovery as an iterative search process that unifies Best-of-N, sequential refinement, tree search, and evolution-style methods under a common compute-allocation view. To isolate allocation effects from prompt engineering and other heuristics, we compare minimal parallel controllers under fixed budgets. On LLM-SRBench equation-discovery tasks, we find that search width is the dominant allocation parameter: the best width in our sweep generally increases with the compute budget, while the population--branching split and controller choice matter less. Appropriate width selection also improves wall-clock efficiency by increasing parallelism. These results suggest that, given an informative verifier, controlling exploration and exploitation is central to scaling LLM-based equation discovery.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-01 in cs.AI).
Categorie arXiv: Computation and Language (cs.CL) ; Artificial Intelligence (cs.AI); Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2608.28660 — 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 2026-10-17T07:28:26.677Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-01?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).