worldbench/awesome-ai-auto-research：🔥 关于人工智能自动研究的调查

项目介绍

worldbench/awesome-ai-auto-research 是一个托管在 GitHub 上的开源项目，主要使用 HTML 编写，采用 MIT License 许可证，目前获得 427 个 Star、32 次 Fork。

仓库全名: worldbench/awesome-ai-auto-research

README 摘要

:sunglasses: Awesome AI Auto-Research :-:| This repository accompanies the survey paper "AI for Auto-Research: Roadmap & User Guide" and tracks papers on AI-assisted and automated scientific research, covering the full research lifecycle. :robot: AI Auto-Research We organize the academic research lifecycle as eight interconnected stages grouped into four epistemological phases. Each phase serves a distinct function…

完整内容见 GitHub 原始 README 或本页。

常见问题

awesome-ai-auto-research 用什么语言开发？

主要使用 HTML 编写，采用 MIT License 许可证，目前有 427 个 Star。

awesome-ai-auto-research 可以免费商用吗？

如何获取 awesome-ai-auto-research 的源码？

可以直接克隆仓库：git clone https://github.com/worldbench/awesome-ai-auto-research.git，默认分支为 main；安装与运行步骤见仓库 README。

awesome-ai-auto-research 最近还在维护吗？

仓库最近一次更新为 7月7日，当前没有未关闭的 Issue；Star 数为 427，可作为社区活跃度的参考。

Visitors

:sunglasses: Awesome AI Auto-Research

This repository accompanies the survey paper “AI for Auto-Research: Roadmap & User Guide” and tracks papers on AI-assisted and automated scientific research, covering the full research lifecycle.

:robot: AI Auto-Research

We organize the academic research lifecycle as eight interconnected stages grouped into four epistemological phases. Each phase serves a distinct function in producing, scrutinizing, and communicating scientific knowledge.


	Phase 1: Creation Generating novel research ideas, searching and synthesizing literature, running coding experiments, and creating publication-quality tables and figures. This phase spans Idea Generation, Literature Review, Coding & Experiments, and Tables & Figures.
	Phase 2: Writing Drafting, editing, and polishing academic manuscripts. AI assistance ranges from semi-automated grammar and citation tools to fully automated paper generation — the most commercially mature yet ethically contested stage.
	Phase 3: Validation Automated peer review generation, reviewer-paper matching, review quality assessment, and AI-assisted author rebuttals. This phase covers Peer Review and Rebuttal & Revision.
	Phase 4: Dissemination Converting papers into slides, posters, videos, websites, and social media content. Each output format targets a different audience and demands its own design logic and AI tool chain.

For additional details, kindly refer to our :books: Paper and :earth_asia: Project Page.

:books: Citation

If you find this work helpful for your research, please kindly consider citing our paper:

@article{survey-ai-auto-research,
  title   = {{AI} for {Auto-Research}: Roadmap \& User Guide},
  author  = {Kong, Lingdong and Sun, Xian and Chow, Wei and Li, Linfeng and Lin, Kevin Qinghong and Zhang, Xuan Billy
             and Wang, Song and Li, Rong and Wu, Qing and Gao, Wei and Wang, Yingshuo and Xie, Shaoyuan
             and Liu, Jiachen and Qu, Leigang and Li, Shijie and Ng, Lai Xing and Cottereau, Benoit R.
             and Liu, Ziwei and Chua, Tat-Seng and Ooi, Wei Tsang},
  journal = {arXiv preprint arXiv:2605.18661},
  year    = {2026}
}

1. Idea Generation
2. Literature Review & Paper Search
3. Coding & Experimentation
4. Tables & Figures
5. Paper Writing
6. Peer Review
7. Rebuttal
8. Dissemination (Paper2X)
9. End-to-End Systems
10. Societal & Critical Perspectives
11. Surveys & Curated Lists
12. Tools & GitHub Repos

1. Idea Generation

LLM Internal Knowledge-Based Generation

Model	Paper	Venue	Website	GitHub

`Chain of Ideas`	Chain of Ideas: Revolutionizing Research Via Novel Idea Development with LLM Agents	arXiv '24	-
`ResearchAgent`	ResearchAgent: Iterative Research Idea Generation over Scientific Literature with Large Language Models	NAACL '25	-
`SciMON`	SciMON: Scientific Inspiration Machines Optimized for Novelty	ACL '24	-
`Idea Gen Agent`	Can LLMs Generate Novel Research Ideas? A Large Scale Human Study with 100+ NLP Researchers	arXiv '24	-	-
`IRIS`	IRIS: Interactive Research Ideation System for Accelerating Scientific Discovery	ACL '25	-
`Spark`	Spark: A System for Scientifically Creative Idea Generation	ICCC '25	-	-
`Diverse Hypo. Search`	Towards Diverse Scientific Hypothesis Search with Large Language Models	arXiv '26	-	-
`Tree-of-Ideas`	Tree-of-Ideas: Automated Research Ideation via Cross-Trajectory Reasoning over Scholarly Evolution	arXiv '26	-	-
`IDEAgent`	IDEAgent: Agentic Quality-Diversity Search for Research Idea Generation	arXiv '26	-	-

External Signal-Driven Generation

Model	Paper	Venue	Website	GitHub
`ActLim-Bench`	Where Research Ideation Begins: Forward-Citation Supervision for Actionable Limitation Generation	arXiv '26	-	-
`IdeaAnchor`	IdeaAnchor: Teaching LLMs to Turn Literature into Research Ideas	arXiv '26	Project	-
`SGHA`	SGHA: Evidence-Grounded Research Problem Discovery with Local Language Models	arXiv '26	-	-
`MAIL`	MAIL: Memory-driven, Adaptive, Incremental, and Literature-grounded Framework for Hypothesis Generation in Chemistry	arXiv '26	-	-

`MOOSE-Chem`	MOOSE-Chem: Large Language Models for Rediscovering Unseen Chemistry Scientific Hypotheses	ICLR '25	-	-
`Nova`	Nova: An Iterative Planning and Search Approach to Enhance Novelty and Diversity of LLM Generated Ideas	arXiv '24	-	-
`SciAgents`	SciAgents: Automating Scientific Discovery through Multi-Agent Intelligent Graph Reasoning	arXiv '24	-
`SciPIP`	SciPIP: An LLM-based Scientific Paper Idea Proposer	arXiv '24	-
`IdeaSynth`	IdeaSynth: Iterative Research Idea Development Through Evolving and Composing Idea Facets with Literature-Grounded Feedback	CHI '25	-	-
`MOOSE-Chem2`	MOOSE-Chem2: Exploring LLM Limits in Fine-Grained Scientific Hypothesis Discovery via Hierarchical Search	NeurIPS '25	-	-
`HALO`	HALO: Interactive Co-abductive Reasoning in Scientific Hypothesis Generation	arXiv '26	-	-
`TCA-SIR`	TCA-SIR: Learning Target-Conditioned Abstractions for Scientific Inspiration Retrieval	arXiv '26	-	-
`ECLAIR`	ECLAIR: A Causally-Grounded AI Framework for Scientific Discovery in Empirical Software Engineering	arXiv '26	-	-
`Principia`	Principia: Test-Time Hypotheses to Principle Construction for Scientific Discovery	arXiv '26	-	Repository (code forthcoming)

Multi-Agent Collaborative Generation

Model	Paper	Venue	Website	GitHub
`HypoForge`	HypoForge: A Self-Improving Multi-Agent Framework for Automated Hypothesis Generation and Testing via Scientific Skill Learning	arXiv '26	-	-

`Combi. Creativity`	Combi. Creativity	arXiv '24	-	-
`Deep Ideation`	Deep Ideation: Designing LLM Agents to Generate Novel Research Ideas on Scientific Concept Network	arXiv '25	-
`VirSci`	Many Heads Are Better Than One: Improved Scientific Idea Generation by A LLM-Based Multi-Agent System	ACL '25	-
`Multi-Agent Dial.`	Multi-Agent Dial.	SIGDIAL '25	-	-
`Artificial Hivemind`	Artificial Hivemind: The Open-Ended Homogeneity of Language Models (and Beyond)	NeurIPS '25	-	-
`Auditable AI Sci.`	Toward Auditable AI Scientists: A Hypothesis Evolution Protocol for LLM Agents	arXiv '26	-	-
`Diverse Personalized Ideation`	Diversifying Personalized Research Ideation against AI-Induced Homogenization	arXiv '26	-	-

Novelty and Feasibility Assessment

Model	Paper	Venue	Website	GitHub
`MechHypoBench`	From Scientific Observations to Mechanisms: Benchmarking Hypothesis Generation by AI Scientists	arXiv '26	-	-
`RATIO`	RATIO: A Benchmark for Retrieval Across Typed Ideation Operations in Scientific Literature	arXiv '26	-	-
`Lit2Test`	What Proves You Wrong: Benchmarking Language Models on Falsifiable Research Ideation	arXiv '26	-	-
`Energy Scoring`	Do LLMs Know a Good Hypothesis When They See One? Logit-Based Energy Scoring Outperforms Prompted LLM-as-Judge for Scientific Hypothesis Ranking	arXiv '26	-	-
`Think-Probe-Respond`	Think-Probe-Respond: Improving Large Language Models as Judges of Research Idea Novelty	EMNLP '26	-	-

`IdeaBench`	IdeaBench: Benchmarking Large Language Models for Research Idea Generation	KDD '25	-	-
`LiveIdeaBench`	Evaluating LLMs’ divergent thinking capabilities for scientific idea generation with minimal context	Nature Communications '26	Website & Leaderboard Data: v1 · DLC-250127 · v2 Models: IdeaWhiz · GGUF
`AI Idea Bench 2025`	AI Idea Bench 2025: AI Research Idea Generation Benchmark	arXiv '25	-
`HeurekaBench`	HeurekaBench: A Benchmarking Framework for AI Co-scientist	ICLR '26	Project / Dataset
`ResearchBench`	ResearchBench: Benchmarking LLMs in Scientific Discovery via Inspiration-Based Task Decomposition	ACL '26	-	-
`HindSight`	HindSight: Evaluating LLM-Generated Research Ideas via Future Impact	arXiv '26	-	-
`Rubric Rewards`	Training AI Co-Scientists Using Rubric Rewards	arXiv '25	-	-
`DeepInnovator`	DeepInnovator: Triggering the Innovative Capabilities of LLMs	arXiv '26	-
`FlowPIE`	FlowPIE: Test-Time Scientific Idea Evolution with Flow-Guided Literature Exploration	arXiv '26	-	-
`SoundnessBench`	SoundnessBench: Can Your AI Scientist Really Tell Good Research Ideas from Bad Ones?	arXiv '26	Project
`LLM-Judge Novelty`	On the Limits of LLM-as-Judge for Scientific Novelty Assessment	arXiv '26	-	-
`LigBench`	LigBench: A Unified and Human-Aligned Benchmark for LLM-based Research Idea Generation	arXiv '26	-	-
`Reconstruction`	Reconstruction: A Blind Benchmark for Recovering Research Ideas from Pre-Publication Bibliographies	arXiv '26	-	-
`AgentIdeaBench`	AgentIdeaBench: Benchmarking Scientific Ideation in the Agent Era	arXiv '26	-
`IdeaAMBIG`	IdeaAMBIG: Benchmarking Implementation-Critical Gaps in Research-Idea Specifications	arXiv '26	-	-
`NovGauge`	NovGauge: A Fine-Grained Benchmark for Diagnosing LLMs’ Capability in Paper Novelty Assessment	arXiv '26	-	-
`Human-AI Ideation Study`	Beyond Better Ideas: How Human-AI Collaboration Shapes Scientific Ideas and Human Perceptions	arXiv '26	-	-
`Tampered ML`	Tampered ML: Evaluating LLM Judgment of Machine Learning Research	arXiv '26	-	Benchmark repository (public split pending)

2. Literature Review & Paper Search

Literature Retrieval

Model	Paper	Venue	Website	GitHub
`ITER`	ITER: Interaction-Aware Retrieval for Agentic Search	arXiv '26	-	-
`Multi-Aspect Retrieval`	Can Retrievers Find the Same Paper from Different Aspects? A Multi-Aspect Full-Paper Scientific Retrieval Benchmark	arXiv '26	-	-

`CiteME`	CiteME: Can Language Models Accurately Cite Scientific Claims?	arXiv '24	-	-
`LitLLM`	LitLLM: A Toolkit for Literature Review with Large Language Models	arXiv '24	-	-
`LitSearch`	LitSearch: A Retrieval Benchmark for Scientific Literature Search	arXiv '24	-
`PaperQA2`	Language Agents Achieve Superhuman Synthesis of Scientific Knowledge	arXiv '24	-
`OpenResearcher`	OpenResearcher: Unleashing AI for Accelerated Scientific Research	EMNLP '24	-	-
`PaSa`	PaSa: An LLM Agent for Comprehensive Academic Paper Search	arXiv '25	-
`Self-Evolving Retrieval`	Towards Self-Evolving Agentic Literature Retrieval	arXiv '26	-	-
`MasterSet`	MasterSet: A Large-Scale Benchmark for Must-Cite Citation Recommendation in the AI/ML Literature	arXiv '26	-	-
`Search, Inspect, Fetch`	Search, Inspect, Fetch: Exploiting Structure-Aware Boolean Retrieval for Deep-Research Agents	arXiv '26	-	-
`Rubric Reranker`	Training Documents Reranker with Search Rubrics for Deep Research Agent	arXiv '26	-	-
`Personalized DR Refinement`	Personalized Deep Research Query Refinement with Graph-Scaffolded Evidence Grounding	arXiv '26	-	-

Model	Paper	Venue	Website	GitHub
`ReviewPilot`	ReviewPilot: A Multi-Agent System for Screening and Data Extraction in Systematic Reviews	arXiv '26	-
`DAS`	Deep Academic Survey: Stateful Agentic Closed-Loop Paradigm for Academic Survey Automation	arXiv '26	Project
`Tree-of-Concerns`	Tree-of-Concerns: Hierarchical Multi-Agent Debate for Unstated-Limitation Extraction in Scientific Critique	EMNLP '26	-	-

`ChatPaper`	ChatPaper: Use LLM to summarize papers	GitHub '23	-
`PaperQA`	PaperQA: Retrieval-Augmented Generative Agent for Scientific Research	arXiv '23	-
`AutoSurvey`	AutoSurvey: Large Language Models Can Automatically Write Surveys	arXiv '24	-
`GPT Researcher`	GPT Researcher: Autonomous Agent for Comprehensive Online Research	GitHub '24	-
`LLMs for Lit. Review`	LLMs for Lit. Review	arXiv '24	-	-
`STORM`	Assisting in Writing Wikipedia-like Articles From Scratch with Large Language Models	arXiv '24	-
`Agentic AutoSurvey`	Agentic AutoSurvey: Let LLMs Survey LLMs	arXiv '25	-	-
`Citegeist`	Citegeist: Automated Generation of Related Work Analysis on the arXiv Corpus	arXiv '25	-	-
`IterSurvey`	IterSurvey: Deep Literature Survey Automation with an Iterative Workflow	arXiv '25	-
`LiRA`	LiRA: A Multi-Agent Framework for Reliable and Readable Literature Review Generation	arXiv '25	-	-
`SurveyForge`	SurveyForge: On the Outline Heuristics, Memory-Driven Generation, and Multi-dimensional Evaluation for Automated Survey Writing	arXiv '25	-
`SurveyG`	SurveyG: A Multi-Agent LLM Framework with Hierarchical Citation Graph for Automated Survey Generation	arXiv '25	-	-
`SurveyX`	SurveyX: Academic Survey Automation via Large Language Models	arXiv '25	-	-
`InteractiveSurvey`	InteractiveSurvey: An LLM-based Personalized and Interactive Survey Paper Generation System	arXiv '25	-
`CiteLLM`	CiteLLM: An Agentic Platform for Trustworthy Scientific Reference Discovery	arXiv '26	-	-
`DeepSurvey`	DeepSurvey: Enhancing Analytical Depth and Citation Reliability in Automated Survey Generation	arXiv '26	-	-
`STRUCTSURVEY`	STRUCTSURVEY: Structured Agentic Retrieval for Automated Survey Paper Generation	arXiv '26	-	-

Deep Research Agents

Model	Paper	Venue	Website	GitHub
`MIRA`	MIRA: A Multimodal Illustrated Deep Research Agent for Controllable Reports	arXiv '26	Project	-
`OpenSaC`	OpenSaC: A Python Library for Search as Code in Deep Research Agents	arXiv '26	-
`HybridResearcher`	HybridResearcher: Bridging the Structured-Unstructured Boundary for Deep Research Agents	arXiv '26	-	-
`JevGPT`	JevGPT: Thinking Fast and Slow in Autonomous Research	arXiv '26	Project
`Crase`	Structurally-bounded Agentic Graph Exploration for Evidence-Grounded Scholarly DeepSearch	arXiv '26	-	-
`DeepWeaver`	DeepWeaver: Bridging the Evidence Synthesis Gap in Open-Ended Question Answering	arXiv '26	-	-
`AgentR`	AgentR A Stateful and Recovery-Aware Software Architecture for LLM-based Auditable Workflows	arXiv '26	-	-

`ASReview`	An Open Source Machine Learning Framework for Efficient and Transparent Systematic Reviews	Nature MI '21	-
`CHIME`	CHIME: LLM-Assisted Hierarchical Organization of Scientific Studies for Literature Review Support	arXiv '24	-	-
`DeepResearch-Agent`	DeepResearchAgent: A Hierarchical Multi-Agent System for Deep Research	GitHub '25	-
`DeerFlow`	DeerFlow: A Deep Research Framework Orchestrating Sub-Agents, Memory, and Sandboxes	GitHub '25	-
`OpenScholar`	OpenScholar: Synthesizing Scientific Literature with Retrieval-Augmented LMs	Nature '26	-	-
`AutoAgent`	AutoAgent	arXiv '25	-	-
`Tongyi DeepResearch`	Tongyi DeepResearch	GitHub '25	-
`O-Researcher`	O-Researcher: An Open Ended Deep Research Model via Multi-Agent Distillation and Agentic RL	arXiv '26	-	-
`OpenResearcher`	OpenResearcher: Unleashing AI for Accelerated Scientific Research	arXiv '26	-
`AREX`	AREX: Towards a Recursively Self-Improving Agent for Deep Research	arXiv '26	-	-
`Predictive Navigation`	Deep Research Pretraining via Predictive Navigation	arXiv '26	-	-
`On-Device DR (4B)`	On-Device Deep Research at 4B: Exposure Bounds Faithfulness, Retrieval Bounds Coverage	arXiv '26	-	-
`Carnot`	Carnot: Interpretable, Interactive, and Optimized Execution of Deep Research Queries	VLDB '26	-	-
`Marginal Value Est.`	Not Worth Another Token: Marginal Value Estimation for Efficient Deep Research Agents	arXiv '26	-	-
`Retrieval-Aware Control`	When Deep Research Agents Stagnate: Enhancing Reasoning with Retrieval-Aware Agent Control	arXiv '26	-	-
`Analogical Deep Research`	Analogical Deep Research: Retrieving and Integrating Historical Analogies for Foresight Analysis	arXiv '26	-	-
`Plato-Bio`	Plato-Bio: verification-first biological novelty screening with temporal rediscovery and structural benchmarks	arXiv '26	-	-
`Albilich`	Albilich: Steerable Proof-State Orchestration for LLM-Based Mathematical Research with CAS Integration	arXiv '26	-	-

Retrieval and Synthesis Quality Assessment

Model	Paper	Venue	Website	GitHub
`IBIS`	From Inertia to Objectivity: Improving Deep Research Agents with Noise Isolation	EMNLP '26	-	-
`Agent to Blame`	Who is the Agent to Blame? Localizing Faithfulness and Citation Mistakes in Agentic Deep Research	EMNLP '26	-	-

`DeepScholar-Bench`	DeepScholar-Bench: A Live Benchmark and Automated Evaluation for Generative Research Synthesis	arXiv '25	-
`ReportBench`	ReportBench: Evaluating Deep Research Agents via Academic Survey Tasks	arXiv '25	-
`IDRBench`	IDRBench: Interactive Deep Research Benchmark	arXiv '26	-	-
`ScholarGym`	ScholarGym: Benchmarking Large Language Model Capabilities in the Information-Gathering Stage of Deep Research	arXiv '26	-	-
`SciNetBench`	SciNetBench: A Relation-Aware Benchmark for Scientific Literature Retrieval Agents	arXiv '26	-	-
`Super Research`	Super Research: Answering Highly Complex Questions with Large Language Models through Super Deep and Super Wide Research	arXiv '26	Website Data: Hugging Face
`AutoResearchBench`	AutoResearchBench: Benchmarking AI Agents on Complex Scientific Literature Discovery	arXiv '26	Website Data: Hugging Face
`PaperMind`	PaperMind: Benchmarking Agentic Reasoning and Critique over Scientific Papers in Multimodal LLMs	arXiv '26	-	-
`DRNOISE`	DRNOISE: Benchmarking Deep Research Agents in Misleading Evidence Environments	arXiv '26	Dataset Evaluation guide
`HiEviDR-Bench`	HiEviDR-Bench: A Benchmark for Hierarchical Evidence Aggregation in Deep Research	arXiv '26	Website Dataset
`SciExplore`	SciExplore: Evaluating Autonomous Agents from Scientific Navigation to Information Integration	arXiv '26	-	-
`WANDR`	WANDR: A Benchmark for Wide and Deep Research	arXiv '26	-	-
`QA-to-DR Bench`	From Simple QA to Deep Research: A Verifiable Benchmark Constructed through Iterative Task Evolution	arXiv '26	-	-
`PRISMA-LLM`	PRISMA-LLM: An Empirical Reporting Framework for AI-Assisted Systematic Reviews	arXiv '26	-	-
`INSPIRE`	Inspire: Benchmarking Scientific Literature Search for Open Research Problems	arXiv '26	-	-
`CESS`	Search Shapes Conclusions: Auditing Evidence Selection Bias in Deep Research Agents	arXiv '26	-	-
`ScholarCatalyst`	ScholarCatalyst: A Benchmark for Retrieving Papers That Inspire New Research	arXiv '26	-
`Deep Research Error Analysis`	Benchmarking and Error Analysis of Deep Search and Deep Research LLM Agents	arXiv '26	-	-
`Real PMIDs, Wrong Papers`	Real PMIDs, Wrong Papers: Harness Design for Biomedical Claim-Verification Agents	arXiv '26	-
`OmniDR-Bench`	Beyond Watching and Searching: Benchmarking Agents for Video-centered Deep Research Reports	arXiv '26	Dataset

3. Coding & Experimentation

Code Generation

Model	Paper	Venue	Website	GitHub

`SWE-bench`	SWE-bench: Can Language Models Resolve Real-World GitHub Issues?	ICLR '24	-
`SWE-agent`	SWE-agent: Agent-Computer Interfaces Enable Automated Software Engineering	arXiv '24	-
`OpenHands`	OpenHands: An Open Platform for AI Software Developers as Generalist Agents	ICLR '25	-
`SWE-bench Pro`	SWE-Bench Pro: Can AI Agents Solve Long-Horizon Software Engineering Tasks?	arXiv '25	-	-
`SWE-EVO`	SWE-EVO: Benchmarking Coding Agents in Long-Horizon Software Evolution Scenarios	arXiv '25	-	-

Paper-to-Code

Model	Paper	Venue	Website	GitHub
`ReproAgent`	ReproAgent: Contract-Guided Paper-to-Code Reproduction	EMNLP '26	-	-
`DeepRepro`	DeepRepro: State-Aware Subplanning for Paper-to-Code Reproduction in Evolving Repositories	CIKM '26	-	-

`FunSearch`	Mathematical Discoveries from Program Search with Large Language Models	Nature '24	-
`SciCode`	SciCode: A Research Coding Benchmark Curated by Scientists	arXiv '24	-
`PaperBench`	PaperBench: Evaluating AI’s Ability to Replicate AI Research	arXiv '25	-
`PaperCoder`	Paper2Code: Automating Code Generation from Scientific Papers in Machine Learning	arXiv '25	-
`ResearchCodeBench`	ResearchCodeBench: Benchmarking LLMs on Implementing Novel ML Research Code	arXiv '25	-	-
`SciReplicate-Bench`	SciReplicate-Bench: Benchmarking LLMs in Agent-driven Algorithmic Reproduction from Research Papers	arXiv '25	-
`PaperCompiler`	PaperCompiler: Faithful Paper-to-Code Generation via Repository-Level Specification Compilation	arXiv '26	-	-

Experiment Execution & Orchestration

Model	Paper	Venue	Website	GitHub
`BMD`	Bayesian Model Discovery with Language Models: Closing the Loop between Hypotheses and Experiments	arXiv '26	-	-
`DAGLoop`	DAGLoop: Aligning Method Design with Experimental Feedback in Automated AI Research	arXiv '26	Project	-
`ResearchTrails`	Learning Scientific Exploration from Human Research Decision Trajectories	arXiv '26	Project	Code
`Praxist`	Praxist: From Experimental Artifacts to Solution Lineages	arXiv '26	-	-
`Skill-Based Baselines`	Can Coding Agents Build Robust Baselines? A Skill-Based Approach for Automating the Medical Imaging Model-Development Pipeline	MICCAI '26	-	-

`BioPlanner`	BioPlanner: Automatic Evaluation of LLMs on Protocol Planning	arXiv '23	-
`CRISPR-GPT`	CRISPR-GPT for Agentic Automation of Gene-Editing Experiments	arXiv '24	-	-
`DS-Agent`	DS-Agent: Automated Data Science by Empowering Large Language Models with Case-Based Reasoning	arXiv '24	-
`MLE-Bench`	MLE-Bench: Evaluating Machine Learning Agents on Machine Learning Engineering	arXiv '24	-	-
`MLAgentBench`	MLAgentBench: Evaluating Language Agents on Machine Learning Experimentation	arXiv '24	-
`MLR-Copilot`	MLR-Copilot: Autonomous Machine Learning Research based on Large Language Models Agents	arXiv '24	-	-
`AIDE`	AIDE: AI-Driven Exploration in the Space of Code	arXiv '25	-
`AlphaEvolve`	AlphaEvolve: A Coding Agent for Scientific and Algorithmic Discovery	arXiv '25	-	-
`AutoReproduce`	AutoReproduce: Automatic AI Experiment Reproduction with Paper Lineage	arXiv '25	-
`CURIE`	Curie: Toward Rigorous and Automated Scientific Experimentation with AI Agents	arXiv '25	-
`MLGym`	MLGym: A New Framework and Benchmark for Advancing AI Research Agents	arXiv '25	-	-
`MLR-Bench`	MLR-Ben