研究索引
论文
从 arXiv、上游 awesome 列表和其他研究源自动收集并分类的论文。
AI CFD科学家:面向开放的计算流体动力学发现,具有物理感知的AI智能体
AI CFD Scientist是一个开源AI智能体,通过物理感知验证自主进行计算流体动力学发现,在壁面Cf RMSE上实现了7.89%的改进。
TruthInsightBench:一个基于证据的基准,用于自动评估开放式科学发现智能体
TruthInsightBench用40项盲任务评测科学发现智能体,以LLM评判证据成熟度;四个编码智能体分数相近,缺乏科学判断。
TruthInsightBench:面向开放式科学发现智能体自动化评估的证据锚定基准
TruthInsightBench 用 40 项盲任务评估科学发现智能体,自动评分显示四类编码智能体停留在平台期,缺乏科学判断。
AutoResearchClaw:基于人机协作的自我强化自主研究
AutoResearchClaw是一个多智能体自主研究流水线,结合人机协作,在ARC-Bench上比AI Scientist v2提升54.7%。
Automating ML for Science: Can Frontier Agents Climb Scientific Hills in the Wild?
Introduces Agent4Sci, a testbed for evaluating frontier agents on scientific ML benchmarks; agents improve results via engineering tricks, not research insight, but plug-in skills boost performance.
SplatLab:面向3DGS研究的自主科学发现
SplatLab autonomously reproduces 3DGS papers, proposes extensions, runs experiments, and writes traceable manuscripts, beating baselines by up to +3.78 dB PSNR.
SurveyGraph: Generating Academic Surveys via Methods-Entity Knowledge Graphs
SurveyGraph generates academic surveys using a Methods-Entity Knowledge Graph, improving retrieval, outline quality, citation support, and semantic relation extraction across evaluations.