{"slug": "discoveryrsi-recursively-learning-to-discover-valuable-hypotheses", "title": "DiscoveryRSI: Recursively Learning to Discover Valuable Hypotheses", "summary": "A paper submitted to arXiv on 16 Oct 2026 introduces DiscoveryRSI, a multi-agent framework that builds hypothesis trees through evolutionary search and recursively self-improves, raising hypothesis quality from 4.86 for the base model to 5.39 after two rounds of self-improvement across four life-science domains. The authors report that their DiscoveryRSI-27B model reached a 6.3% rate of hypothesis correspondence with research published after the model cutoff, versus 2.5% for GPT-6 and 2.3% for Claude Opus 5, and scored higher than GPT-6 Luna in preliminary human assessments of hypothesis importance and evidence-bounded plausibility. The paper also introduces DFM-Bench, which assesses hypothesis quality and cumulative research progress at the levels of hypotheses, revisions, development paths and research runs.", "body_md": "# Computer Science > Computation and Language\n\n  [Submitted on 16 Oct 2026]\n\n# Title:DiscoveryRSI: Recursively Learning to Discover Valuable Hypotheses\n\n[View PDF](http://arxiv.org/pdf/2610.23839v1)\n\n[HTML (experimental)](https://arxiv.org/html/2610.23839v1)\n\nAbstract:Scientific discovery requires developing worthwhile hypotheses into practical investigations, yet learning from this evolving research process remains challenging. We introduce an operational definition of valuable and testable research hypotheses: evidence-grounded claims about consequential uncertainties, coupled with feasible investigations, explicit decision criteria, and alternative explanations. Our DiscoveryRSI framework develops hypothesis trees through multi-agent evolutionary search, compiles selected investigations into executable Harbor tasks, and recursively learns through three mechanisms: hypothesis credit optimization attributes scientific assessments to documented research contributions; harness replay identifies productive strategies and recurring failures to refine research workflows; and rediscovery applies accepted improvements to new investigations, generating further experience for subsequent learning. We also introduce DFM-Bench, which provides comprehensive assessment of hypothesis quality and cumulative research progress at the levels of hypotheses, revisions, development paths and research runs. Across four life-science domains, automated evaluation shows that hypothesis quality rises from 4.86 for the base model to 5.39 after two rounds of self-improvement, with the large gains in falsifiability and data resolvability. We further validate hypothesis quality through expert review, comparing our DiscoveryRSI-27B with other research agents. Preliminary human assessments show that DiscoveryRSI-27B scores higher than GPT-6 Luna on hypothesis importance and whether claims are plausible and evidence-bounded. DiscoveryRSI-27B achieved a 6.3% rate of hypothesis correspondence with research published after the model cutoff, largely outperform GPT-6 with 2.5% and Claude Opus 5 with 2.3%, while delivering greater human-rated improvement along research paths.\n    \n\n### Additional Features\n\n### References & Citations\n\nLoading...\n\n# Bibliographic and Citation Tools\n\nBibliographic Explorer \n\n*(*[What is the Explorer?](https://info.arxiv.org/labs/showcase.html#arxiv-bibliographic-explorer))\nConnected Papers \n\n*(*[What is Connected Papers?](https://www.connectedpapers.com/about))\nLitmaps \n\n*(*[What is Litmaps?](https://www.litmaps.co/))\nscite Smart Citations \n\n*(*[What are Smart Citations?](https://www.scite.ai/))\n# Code, Data and Media Associated with this Article\n\nalphaXiv \n\n*(*[What is alphaXiv?](https://alphaxiv.org/))\nCatalyzeX Code Finder for Papers \n\n*(*[What is CatalyzeX?](https://www.catalyzex.com))\nDagsHub \n\n*(*[What is DagsHub?](https://dagshub.com/))\nGotit.pub \n\n*(*[What is GotitPub?](http://gotit.pub/faq))\nHugging Face \n\n*(*[What is Huggingface?](https://huggingface.co/huggingface))\nScienceCast \n\n*(*[What is ScienceCast?](https://sciencecast.org/welcome))\n# Demos\n\n# Recommenders and Search Tools\n\nInfluence Flower \n\n*(*[What are Influence Flowers?](https://influencemap.cmlab.dev/))\nCORE Recommender \n\n*(*[What is CORE?](https://core.ac.uk/services/recommender))\n# arXivLabs: experimental projects with community collaborators\n\narXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.\n\nBoth individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.\n\nHave an idea for a project that will add value for arXiv's community? [**Learn more about arXivLabs**](https://info.arxiv.org/labs/index.html).", "url": "https://wpnews.pro/news/discoveryrsi-recursively-learning-to-discover-valuable-hypotheses", "canonical_source": "http://arxiv.org/abs/2610.23839v1", "published_at": "2026-10-19 14:35:27+00:00", "updated_at": "2026-10-19 15:17:28.814241+00:00", "lang": "en", "topics": ["ai-research", "ai-agents", "large-language-models", "artificial-intelligence"], "entities": ["DiscoveryRSI", "DiscoveryRSI-27B", "DFM-Bench", "GPT-6 Luna", "Claude Opus 5", "arXiv"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/discoveryrsi-recursively-learning-to-discover-valuable-hypotheses", "markdown": "https://wpnews.pro/news/discoveryrsi-recursively-learning-to-discover-valuable-hypotheses.md", "text": "https://wpnews.pro/news/discoveryrsi-recursively-learning-to-discover-valuable-hypotheses.txt", "jsonld": "https://wpnews.pro/news/discoveryrsi-recursively-learning-to-discover-valuable-hypotheses.jsonld"}}