TWITTER / X
论文 DiscoveryRSI：将递归自我改进用于科学假设生成，27B 模型假设与后续已发表研究匹配率称高于 GPT-6
账号 @omarsar0 转发介绍论文 DiscoveryRSI：该系统将递归自我改进（RSI）应用于四个生命科学领域的科学假设生成。据帖子作者转述，其 27B 模型生成的假设与训练截止后发表的研究相匹配的比例为 6.3%，高于 GPT-6 的 2.5% 和 Claude Opus 5 的 2.3%；系统通过多智能体进化搜索扩展假设树、将研究转化为可执行任务，并在每轮后回放学到的策略进行迭代，自动假设质量评分经两轮从 4.86 升至 5.39，可证伪性提升最大。论文见 arXiv:2610.23839。
AI 解读
该帖由 @omarsar0 转发，介绍一篇关于递归自我改进（RSI）在科学假设生成中应用的论文。原文明确陈述的核心内容：系统名为 DiscoveryRSI，将 RSI 应用于四个生命科学领域的假设生成；采用多智能体进化搜索构建假设树，将选中的研究转化为可执行任务，并在每轮运行后为有贡献的假设分配信用、回放研究流程以总结有效策略与重复失败，再迭代改进；质量评分在两轮迭代后从 4.86 升至 5.39，可证伪性提升最大。需要区分的是：27B 模型假设与训练截止后发表的论文匹配率为 6.3%、GPT-6 为 2.5%、Claude Opus 5 为 2.3% 等数字均来自帖子作者对论文的转述（"match research published after its training cutoff"的表述本身值得注意，需注意这不等同于假设被验证正确，而是与后续发表研究的匹配），属于作者声称的论文结果，未经独立核实。论文链接指向 arXiv（abs/2610.23839），并附 dair.ai 的"Chat with Paper"入口。行业影响方面（推断）：该工作属于"AI for Science"与多智能体自我改进的交叉方向，若其评估方法可靠，对小模型在科研假设生成上的能力上限是一个有信息量的数据点，但其"匹配率"指标的定义与统计口径需要阅读论文原文才能判断其意义。
推文原文
RT elvis Recursive self-improvement is one of the most discussed ideas in AI right now. I am particularly interested in how it applies to science. I saw this really cool new paper with neat ideas, especially on hypothesis generation. A good read. Specifically, they propose DiscoveryRSI, which applies RSI to scientific hypothesis generation in four life-science domains. Its 27B model produced hypotheses that match research published after its training cutoff 6.3% of the time, compared with 2.5% for GPT-6 and 2.3% for Claude Opus 5. The system grows trees of hypotheses with multi-agent evolutionary search and turns the chosen investigations into executable tasks. After each run, it assigns credit to the hypotheses that contributed, replays the research harness to find productive strategies and repeated failures, and applies accepted improvements to new investigations. Automated hypothesis quality rose from 4.86 to 5.39 after two rounds, with the largest gains in falsifiability. https://arxiv.org/abs/2610.23839 Chat with Paper: https://academy.dair.ai/papers/discoveryrsi-recursively-learning-to-discover-valuable-hypotheses-2610.23839