mlsanityData to 7 Oct 2026

DeepScientist: Advancing Frontier-Pushing Scientific Findings Progressively

arXiv 2509.26603 ↗First posted 30 Sep 2025cs.CL cs.LGcited by 56

Not on the map. Why?

Cited by 32 in LLM agents and tool use, but placed by its own abstract, which didn’t fit.

Abstract (arXiv)

While previous AI Scientist systems can generate novel findings, they often lack the focus to produce scientifically valuable contributions that address pressing human-defined challenges. We introduce DeepScientist, a system designed to overcome this by conducting goal-oriented, fully autonomous scientific discovery over month-long timelines. It formalizes discovery as a Bayesian Optimization problem, operationalized through a hierarchical evaluation process consisting of "hypothesize, verify, and analyze". Leveraging a cumulative Findings Memory, this loop intelligently balances the exploration of novel hypotheses with exploitation, selectively promoting the most promising findings to higher-fidelity levels of validation. Consuming over 20,000 GPU hours, the system generated about 5,000 unique scientific ideas and experimentally validated approximately 1100 of them, ultimately surpassing human-designed state-of-the-art (SOTA) methods on three frontier AI tasks by 183.7%, 1.9%, and 7.9%. This work provides the first large-scale evidence of an AI achieving discoveries that progressively surpass human SOTA on scientific tasks, producing valuable findings that genuinely push the frontier of scientific discovery. To facilitate further research into this process, we will open-source all experimental logs and system code at https://github.com/ResearAI/DeepScientist/.

Key facts. Cited by 56 papers in the map (data through 7 Oct 2026). Builds on 42 of its 55 references that are in the map. First posted on arXiv 30 Sep 2025.

Builds on

42 of 55 references.

References · 42

  1. 2023Efficient Memory Management for Large Language Model Serving with PagedAttentioncited by 5,767
  2. 2021Program Synthesis with Large Language Modelscited by 3,509
  3. 2024The AI Scientist: Towards Fully Automated Open-Ended Scientific Discoverycited by 1,002
  4. 2025AlphaEvolve: A coding agent for scientific and algorithmic discoverycited by 763
  5. 2018A Tutorial on Bayesian Optimizationcited by 734
  6. 2023The Falcon Series of Open Language Modelscited by 526
  7. 2024MLE-bench: Evaluating Machine Learning Agents on Machine Learning Engineeringcited by 429
  8. 2025Agent Laboratory: Using LLM Agents as Research Assistantscited by 405
  9. 2025The AI Scientist-v2: Workshop-Level Automated Scientific Discovery via Agentic Tree Searchcited by 375
  10. 2025Accelerating scientific discovery with Co-Scientistcited by 353
  11. 2022Large Language Models are Better Reasoners with Self-Verificationcited by 304
  12. 2025PaperBench: Evaluating AI's Ability to Replicate AI Researchcited by 286
30 more
  1. 2023Fast-DetectGPT: Efficient Zero-Shot Detection of Machine-Generated Text via Conditional Probability Curvaturecited by 278
  2. 2024Unlocking Efficiency in Large Language Model Inference: A Comprehensive Survey of Speculative Decodingcited by 266
  3. 2025Model Context Protocol (MCP): Landscape, Security Threats, and Future Research Directionscited by 257
  4. 2024Spotting LLMs With Binoculars: Zero-Shot Detection of Machine-Generated Textcited by 255
  5. 2023DetectLLM: Leveraging Log Rank Information for Zero-Shot Detection of Machine-Generated Textcited by 207
  6. 2024ResearchAgent: Iterative Research Idea Generation over Scientific Literature with Large Language Modelscited by 200
  7. 2023RADAR: Robust AI-Text Detection via Adversarial Learningcited by 187
  8. 2025Which Agent Causes Task Failures and When? On Automated Failure Attribution of LLM Multi-Agent Systemscited by 170
  9. 2024RAID: A Shared Benchmark for Robust Evaluation of Machine-Generated Text Detectorscited by 157
  10. 2025ShinkaEvolve: Towards Open-Ended And Sample-Efficient Program Evolutioncited by 146
  11. 2024CycleResearcher: Improving Automated Research via Automated Reviewcited by 142
  12. 2025AgenTracer: Who Is Inducing Failure in the LLM Agentic Systems?cited by 105
  13. 2025DeepReview: Improving LLM-based Paper Review with Human-like Deep Thinking Processcited by 97
  14. 2025AgentRxiv: Towards Collaborative Autonomous Researchcited by 64
  15. 2025An AI system to help scientists write expert-level empirical softwarecited by 52
  16. 2024Personality Alignment of Large Language Modelscited by 36
  17. 2025An Empirical Analysis of Uncertainty in Large Language Model Evaluationscited by 20
  18. 2025AI Scientists Fail Without Strong Implementation Capabilitycited by 20
  19. 2025Paper2Agent: Reimagining Research Papers As Interactive and Reliable AI Agentscited by 19
  20. 2023Evolving Scientific Discovery by Unifying Data and Background Knowledge with AI Hilbertcited by 18
  21. 2024ControlLM: Crafting Diverse Personalities for Language Modelscited by 15
  22. 2025How Far Are AI Scientists from Changing the World?cited by 15
  23. 2024Memory Is All You Need: An Overview of Compute-in-Memory Architectures for Accelerating Large Language Model Inferencecited by 14
  24. 2024Glimpse: Enabling White-Box Methods to Use Proprietary Models for Zero-Shot LLM-Generated Text Detectioncited by 14
  25. 2023Mastering Symbolic Operations: Augmenting Language Models with Compiled Neural Networkscited by 12
  26. 2024Does Knowledge Localization Hold True? Surprising Differences Between Entity and Relation Perspectives in Language Modelscited by 8
  27. 2025Scaling Laws in Scientific Discovery with AI and Robot Scientistscited by 5
  28. 2025The Need for Verification in AI-Driven Scientific Discoverycited by 5
  29. 2022Artificial Text Detection with Multiple Training Strategiescited by 4
  30. 2025AutomataGPT: Forecasting and Ruleset Inference for Two-Dimensional Cellular Automatacited by 2

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 8; 2026: 48.

56 papers.

Sort:

By area:

  1. 2025Memory in the Age of AI Agentscited by 285
  2. 2026EvoScientist: Towards Multi-Agent Evolving AI Scientists for End-to-End Scientific Discoverycited by 41
  3. 2026MARS: Modular Agent with Reflective Search for Automated AI Researchcited by 22
  4. 2026AutoFigure: Generating and Refining Publication-Ready Scientific Illustrationscited by 20
  5. 2026AutoSOTA: An End-to-End Automated Research System for State-of-the-Art AI Model Discoverycited by 17
  6. 2025OmniScientist: Toward a Co-evolving Ecosystem of Human and AI Scientistscited by 15
  7. 2026ResearchGym: Evaluating Language Model Agents on Real-World AI Researchcited by 15
  8. 2026Toward Autonomous Long-Horizon Engineering for ML Researchcited by 15
  9. 2026ScientistOne: Towards Human-Level Autonomous Research via Chain-of-Evidencecited by 15
  10. 2025Jr. AI Scientist and Its Risk Report: Autonomous Scientific Exploration from a Baseline Papercited by 10
  11. 2026AutoResearchBench: Benchmarking AI Agents on Complex Scientific Literature Discoverycited by 10
  12. 2025A Survey of AI Scientistscited by 9
38 more
  1. 2026SciCoQA: Quality Assurance for Scientific Paper--Code Alignmentcited by 9
  2. 2026AutoResearch AI: Towards AI-Powered Research Automation for Scientific Discoverycited by 9
  3. 2026NatureBench: Can Coding Agents Match the Published SOTA of Nature-Family Papers?cited by 9
  4. 2026AI Can Learn Scientific Tastecited by 8
  5. 2026DeepReviewer 2.0: A Traceable Agentic System for Auditable Scientific Peer Reviewcited by 7
  6. 2026InnoEval: On Research Idea Evaluation as a Knowledge-Grounded, Multi-Perspective Reasoning Problemcited by 6
  7. 2026TREX: Automating LLM Fine-tuning via Agent-Driven Tree-based Explorationcited by 6
  8. 2025Deep Ideation: Designing LLM Agents to Generate Novel Research Ideas on Scientific Concept Networkcited by 5
  9. 2026Towards a Medical AI Scientistcited by 5
  10. 2026FARS: A Fully Automated Research System Deployed at Scalecited by 5
  11. 2026DeepInnovator: Triggering the Innovative Capabilities of LLMscited by 4
  12. 2026Visual Latents Know More Than They Say: Unsilencing Latent Reasoning in MLLMscited by 4
  13. 2026NanoResearch: Co-Evolving Skills, Memory, and Policy for Personalized Research Automationcited by 4
  14. 2026AutoSci: A Memory-Centric Agentic System for the Full Scientific Research Lifecyclecited by 4
  15. 2026Let the Abyss Stare Back Adaptive Falsification for Autonomous Scientific Discoverycited by 3
  16. 2026AI CFD Scientist: Toward Open-Ended Computational Fluid Dynamics Discovery with Physics-Aware AI Agentscited by 3
  17. 2026Federation over Text: Insight Sharing for Multi-Agent Reasoningcited by 2
  18. 2026Toward AI VIS Co-Scientists: A General and End-to-End Agent Harness for Solving Complex Data Visualization Taskscited by 2
  19. 2025Structural Enforcement of Statistical Rigor in AI-Driven Discovery: A Functional Architecturecited by 1
  20. 2025HybridQuestion: Human-AI Collaboration for Identifying High-Impact Research Questionscited by 1
  21. 2026BioProAgent: Neuro-Symbolic Grounding for Constrained Scientific Planningcited by 1
  22. 2026Embodied Science: Closing the Discovery Loop with Agentic Embodied AIcited by 1
  23. 2026SciResearcher: Scaling Deep Research Agents for Frontier Scientific Reasoningcited by 1
  24. 2026Towards Characterizing Scientific Image Utility and Upgradabilitycited by 1
  25. 2026StatefulDiscovery: Evidence-Calibrated Claim Formation in Open-Ended Scientific Discoverycited by 1
  26. 2026One Reflection Is Not Enough: Self-Correcting Autonomous Research via Multi-Hypothesis Failure Attributioncited by 1
  27. 2026Towards Autonomous and Auditable Medical Imaging Model Developmentcited by 1
  28. 2026ReasFlow: Assisting Reasoning-Centric Scientific Discovery in Applied Mathematics via a Knowledge-Based Multi-Agent Systemcited by 1
  29. 2026Can LLM Agents Discover? Evaluating Creativity on ML Engineering Taskscited by 1
  30. 2025CodeDistiller: Automatically Generating Code Libraries for Scientific Coding Agentscited by 0
  31. 2026EDM-ARS: A Domain-Specific Multi-Agent System for Automated Educational Data Mining Researchcited by 0
  32. 2026Story2Proposal: A Scaffold for Structured Scientific Paper Writingcited by 0
  33. 2026Paper Reconstruction Evaluation: Evaluating Presentation and Hallucination in AI-written Paperscited by 0
  34. 2026AgentEconomist: An End-to-end Agentic System Translating Economic Intuitions into Executable Computational Experimentscited by 0
  35. 2026LECTOR: Joint Optimization of Scientific Reasoning Graphs and Introduction Generationcited by 0
  36. 2026ARAC: Benchmarking Auto-Research's Alignment and Completeness on End-to-End Researchscited by 0
  37. 2026Position: AI Agents in Scientific Teams Should Be Studied as Human-Agent Systemscited by 0
  38. 2026PaperGym: Rubric-Centered Evolution for Research-Plan Generationcited by 0

50 most cited of 56 All 56 (JSON)