将机器学习与生物学结合，构建可复用的模型、流程与数据集，服务表观遗传、衰老与癌症研究。 Bringing machine learning to biology — reusable models, pipelines, and datasets for epigenetics, aging, and cancer research.
Claude + Kimi + Codex 三席多智能体发散问答台：带着科学/技术问题来，三席各开一个角度、相互启发，Claude 摊开成"可能性地图"，多轮追问。 A three-seat multi-agent discussion desk (Claude + Kimi + Codex). Bring a science/tech question — each seat opens a different angle, they cross-pollinate, and Claude lays out a "possibility map" over multiple rounds.
面向文献/证据的迭代式发散搜索：自动多轮联网检索、扩展查询、按新颖度筛选，汇总成可追溯的发现地图。 Iterative divergent literature/evidence search: automated multi-round web retrieval, query expansion, and novelty-filtered synthesis into a traceable map of findings.
注：两个工具需登录（账号 taolab）。服务部署在本课题组服务器上。 Note: both tools require sign-in (user: taolab). Hosted on the lab's own server.
以下是课题组在日常科研中反复实践、也反复踩坑之后总结的九条经验。它们不是规定，而是我们目前的共识，会随着实践继续修正。 Nine lessons the lab has arrived at through repeated practice — and repeated mistakes. They are not rules but our current consensus, and they will keep being revised as we go.