TWITTER / X
DiscoveryRSI：将递归自我改进用于科学假设生成，27B 模型表现超过 GPT-6 与 Claude Opus 5
@omarsar0 转发介绍一篇新论文：DiscoveryRSI 将递归自我改进（RSI）引入科学假设生成，覆盖生命科学四个领域。其 27B 模型生成的假设有 6.3% 与训练截止后发表的论文吻合（GPT-6 为 2.5%，Claude Opus 5 为 2.3%）。系统通过多智能体进化搜索构建假设树，将选定的研究方向转化为可执行任务，并在每轮运行后对有贡献的假设做信用分配、回放研究流程以提炼有效策略。两轮迭代后，自动化假设质量从 4.86 升至 5.39，可证伪性提升最大。论文：https://arxiv.org/abs/2610.23839
AI 解读
该帖作者（转发 elvis 的内容）介绍了一篇关于递归自我改进（RSI）在科学假设生成中应用的新论文。原文明确陈述的要点：1) 论文提出 DiscoveryRSI，将 RSI 应用于生命科学四个领域的科学假设生成；2) 其 27B 模型生成的假设有 6.3% 与研究截止点之后发表的论文吻合，高于 GPT-6 的 2.5% 和 Claude Opus 5 的 2.3%；3) 系统通过多智能体进化搜索扩展假设树，并将选定的研究方向转化为可执行任务；4) 每次运行后系统会回溯对有贡献的假设进行信用分配，重放研究流程以发现有效策略与重复失败，并将接受的改进应用于新研究；5) 经过两轮迭代，自动化假设质量评分从 4.86 升至 5.39，其中可证伪性提升最大。以上均为作者转述的论文结论，未经独立验证。新闻价值判断：这是 AI for Science / 智能体方向的具体研究进展，含可核查的论文链接与量化结果，符合本订阅范围，属可发布内容。行业影响（推断部分）：若此类 RSI 驱动的假设生成方法被后续工作复现，可能加速生命科学领域的自动化研究流程，并对评估 AI 科学产出的基准方法提出新思路。
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