Difference between revisions of "Science Agents"

From GISAXS
Jump to: navigation, search
(→‎AI Science Systems)
(→‎(Pre) Generate Articles)
 
(16 intermediate revisions by the same user not shown)
Line 4: Line 4:
 
==Literature==
 
==Literature==
 
* [https://www.alphaxiv.org/explore alphaXiv | Explore]: Understand arXiv papers
 
* [https://www.alphaxiv.org/explore alphaXiv | Explore]: Understand arXiv papers
 +
* 2026-02: [https://www.nature.com/articles/s41586-025-10072-4 Synthesizing scientific literature with retrieval-augmented language models]
  
 
===LLM extract data from papers===
 
===LLM extract data from papers===
Line 20: Line 21:
 
* 2024-02: [https://arxiv.org/abs/2408.07055 LongWriter: Unleashing 10,000+ Word Generation from Long Context LLMs] ([https://github.com/THUDM/LongWriter code])
 
* 2024-02: [https://arxiv.org/abs/2408.07055 LongWriter: Unleashing 10,000+ Word Generation from Long Context LLMs] ([https://github.com/THUDM/LongWriter code])
 
* 2024-08: Scientific papers: [https://arxiv.org/abs/2408.06292 The AI Scientist: Towards Fully Automated Open-Ended Scientific Discovery]
 
* 2024-08: Scientific papers: [https://arxiv.org/abs/2408.06292 The AI Scientist: Towards Fully Automated Open-Ended Scientific Discovery]
 +
** 2026-04: [https://www.nature.com/articles/s41586-026-10265-5 Towards end-to-end automation of AI research]
 
* 2024-09: PaperQA2: [https://paper.wikicrow.ai/ Language Models Achieve Superhuman Synthesis of Scientific Knowledge] ([https://x.com/SGRodriques/status/1833908643856818443 𝕏 post], [https://github.com/Future-House/paper-qa code])
 
* 2024-09: PaperQA2: [https://paper.wikicrow.ai/ Language Models Achieve Superhuman Synthesis of Scientific Knowledge] ([https://x.com/SGRodriques/status/1833908643856818443 𝕏 post], [https://github.com/Future-House/paper-qa code])
 
* 2025-03: [https://arxiv.org/abs/2503.18866 Reasoning to Learn from Latent Thoughts]
 
* 2025-03: [https://arxiv.org/abs/2503.18866 Reasoning to Learn from Latent Thoughts]
Line 39: Line 41:
 
* 2025-06: [https://arxiv.org/abs/2506.00794 Predicting Empirical AI Research Outcomes with Language Models]
 
* 2025-06: [https://arxiv.org/abs/2506.00794 Predicting Empirical AI Research Outcomes with Language Models]
 
* 2025-06: [https://arxiv.org/abs/2506.20803 The Ideation-Execution Gap: Execution Outcomes of LLM-Generated versus Human Research Ideas]
 
* 2025-06: [https://arxiv.org/abs/2506.20803 The Ideation-Execution Gap: Execution Outcomes of LLM-Generated versus Human Research Ideas]
 +
* 2026-03: [https://arxiv.org/abs/2603.14473 AI Can Learn Scientific Taste]
  
 
==Adapting LLMs to Science==
 
==Adapting LLMs to Science==
Line 79: Line 82:
 
* [https://github.com/TheBlewish/Automated-AI-Web-Researcher-Ollama Automated-AI-Web-Researcher-Ollama]
 
* [https://github.com/TheBlewish/Automated-AI-Web-Researcher-Ollama Automated-AI-Web-Researcher-Ollama]
 
* 2025-01: [https://arxiv.org/abs/2501.05366 Search-o1: Agentic Search-Enhanced Large Reasoning Models] ([https://search-o1.github.io/ project], [https://github.com/sunnynexus/Search-o1 code])
 
* 2025-01: [https://arxiv.org/abs/2501.05366 Search-o1: Agentic Search-Enhanced Large Reasoning Models] ([https://search-o1.github.io/ project], [https://github.com/sunnynexus/Search-o1 code])
 +
* 2026-02: [https://www.nature.com/articles/s41586-025-10072-4 Synthesizing scientific literature with retrieval-augmented language models] ([https://allenai.org/blog/openscholar-nature blog])
  
 
===Commercial===
 
===Commercial===
Line 88: Line 92:
 
* [https://periodic.com/ Periodic Labs]
 
* [https://periodic.com/ Periodic Labs]
 
* [https://edisonscientific.com/articles/announcing-edison-scientific Edison Scientific] (drug discovery, spinoff from [https://www.futurehouse.org/ FutureHouse])
 
* [https://edisonscientific.com/articles/announcing-edison-scientific Edison Scientific] (drug discovery, spinoff from [https://www.futurehouse.org/ FutureHouse])
 +
* 2026-03: Mirendil Inc.: advanced models to speed up R&D in scientific domains, especially biology and materials science
  
 
====Bio====
 
====Bio====
Line 196: Line 201:
 
* 2025-11: [https://arxiv.org/abs/2511.08151 SciAgent: A Unified Multi-Agent System for Generalistic Scientific Reasoning]
 
* 2025-11: [https://arxiv.org/abs/2511.08151 SciAgent: A Unified Multi-Agent System for Generalistic Scientific Reasoning]
 
* 2026-02: [https://arxiv.org/abs/2601.23265 PaperBanana: Automating Academic Illustration for AI Scientists]
 
* 2026-02: [https://arxiv.org/abs/2601.23265 PaperBanana: Automating Academic Illustration for AI Scientists]
 +
* 2026-03: [https://arxiv.org/abs/2603.20179 AI Agents Can Already Autonomously Perform Experimental High Energy Physics]
  
 
==Science Multi-Agent Setups==
 
==Science Multi-Agent Setups==
 
* 2025-01: [https://arxiv.org/abs/2501.04227 Agent Laboratory: Using LLM Agents as Research Assistants]
 
* 2025-01: [https://arxiv.org/abs/2501.04227 Agent Laboratory: Using LLM Agents as Research Assistants]
 
* 2025-04: [https://www.nature.com/articles/s41551-025-01363-2 Coordinated AI agents for advancing healthcare] ([https://www.nature.com/articles/s41551-025-01363-2.epdf?sharing_token=CIYP3J8LZE4BX31fV3WxUdRgN0jAjWel9jnR3ZoTv0O9iD-yhgqzRaz_7VASayWRePPhWDD2xFyfuOpSXbdPaOtt7oH4nfXo7telALzNwY3V1p9SxoqBEJy2OuaJ_cA35-CYQC1XgjCNTZUw46dh1KX-Dj8e7-1Vk_RlZKFLrc8%3D pdf])
 
* 2025-04: [https://www.nature.com/articles/s41551-025-01363-2 Coordinated AI agents for advancing healthcare] ([https://www.nature.com/articles/s41551-025-01363-2.epdf?sharing_token=CIYP3J8LZE4BX31fV3WxUdRgN0jAjWel9jnR3ZoTv0O9iD-yhgqzRaz_7VASayWRePPhWDD2xFyfuOpSXbdPaOtt7oH4nfXo7telALzNwY3V1p9SxoqBEJy2OuaJ_cA35-CYQC1XgjCNTZUw46dh1KX-Dj8e7-1Vk_RlZKFLrc8%3D pdf])
 +
 +
=Science Agentic Components=
 +
==Frameworks==
 +
* [https://platform.claude.com/docs/en/agent-sdk/overview Anthropic Claude Agent SKD overview]
 +
* [https://openclaw.ai/ OpenClaw]
 +
* [https://opencode.ai/ OpenCode]
 +
* [https://github.com/OpenHands/software-agent-sdk OpenHands]
 +
* [https://github.com/lamm-mit?tab=repositories LAMM: MIT Laboratory for Atomistic and Molecular Mechanics]
 +
** [https://github.com/lamm-mit/scienceclaw ScienceClaw]: Framework for autonomous scientific investigation without central coordination.
 +
** [https://infinite-lamm.vercel.app/ Infinite]: The Infinite Corridor of Scientific Discovery. Open science, powered by many — agents and humans discovering together.
 +
 +
==Personalities==
 +
* 2026-03: [https://github.com/msitarzewski/agency-agents The Agency: AI Specialists Ready to Transform Your Workflow]
 +
 +
==Skills==
 +
* 2026-03: [https://github.com/K-Dense-AI/claude-scientific-skills/tree/main?tab=readme-ov-file#use-cases Claude Scientific Skills] (list)
  
 
=AI Science Systems=
 
=AI Science Systems=
Line 288: Line 310:
 
** 2026-01: Google DeepMind: [https://arxiv.org/abs/2601.22401 Semi-Autonomous Mathematics Discovery with Gemini: A Case Study on the Erdős Problems]
 
** 2026-01: Google DeepMind: [https://arxiv.org/abs/2601.22401 Semi-Autonomous Mathematics Discovery with Gemini: A Case Study on the Erdős Problems]
 
*** Attempted 700 problems, solved 13 open Erdős problems: 5 novel autonomous solutions, 8 through existing literature.
 
*** Attempted 700 problems, solved 13 open Erdős problems: 5 novel autonomous solutions, 8 through existing literature.
 +
** 2026-02: [https://www.erdosproblems.com/846 Erdős Problem #846]
 +
*** [https://x.com/roydanroy/status/2026804567178953048?s=20 Google DeepMind]
 +
*** [https://x.com/mehtaab_sawhney/status/2026716221933343147?s=20 Using OpenAI internal model] (paper: [https://cdn.openai.com/infinite-sets/main_single_clean3.pdf On infinite sets with no 3 on a line])
 +
** 2026-03: Three problems solved using OpenAI GPT internal model. Paper: [https://arxiv.org/pdf/2603.29961 Short Proofs in Combinatorics and Number Theory]
 
* 2026-01: [https://arxiv.org/abs/2601.07222 The motivic class of the space of genus 0 maps to the flag variety]
 
* 2026-01: [https://arxiv.org/abs/2601.07222 The motivic class of the space of genus 0 maps to the flag variety]
 
* 2026-02: Google DeepMind: [https://arxiv.org/abs/2602.10177 Towards Autonomous Mathematics Research]
 
* 2026-02: Google DeepMind: [https://arxiv.org/abs/2602.10177 Towards Autonomous Mathematics Research]
 +
* 2026-03: Donald Knuth: [https://www-cs-faculty.stanford.edu/~knuth/papers/claude-cycles.pdf A problem in Directed Hamiltonian Cycles] solved by Filip Stappers using Claude Opus 4.6
 +
* 2026-03: Google DeepMind: [https://arxiv.org/abs/2603.09172 Reinforced Generation of Combinatorial Structures: Ramsey Numbers]
 +
* 2026-03: [https://epoch.ai/frontiermath/open-problems FrontierMath] problem: [https://epoch.ai/frontiermath/open-problems/ramsey-hypergraphs "A Ramsey-style Problem on Hypergraphs"] solved by Kevin Barreto and Liam Price using GPT-5.4 Pro
  
 
==Physics assistance==
 
==Physics assistance==
Line 295: Line 324:
 
* 2025-12: [https://www.sciencedirect.com/science/article/pii/S0370269325008111 Relativistic covariance and nonlinear quantum mechanics: Tomonaga-Schwinger analysis]
 
* 2025-12: [https://www.sciencedirect.com/science/article/pii/S0370269325008111 Relativistic covariance and nonlinear quantum mechanics: Tomonaga-Schwinger analysis]
 
** [https://x.com/hsu_steve/status/1996034522308026435?s=20 Steve Hsu], [https://drive.google.com/file/d/16sxJuwsHoi-fvTFbri9Bu8B9bqA6lr1H/view Theoretical Physics with Generative AI]
 
** [https://x.com/hsu_steve/status/1996034522308026435?s=20 Steve Hsu], [https://drive.google.com/file/d/16sxJuwsHoi-fvTFbri9Bu8B9bqA6lr1H/view Theoretical Physics with Generative AI]
 +
* 2026-02: [https://arxiv.org/abs/2602.12176 Single-minus gluon tree amplitudes are nonzero] (GPT-5.2, [https://openai.com/index/new-result-theoretical-physics/ blog])
 +
 
==Literature exploration==
 
==Literature exploration==
 
* 2025-11: [https://arxiv.org/abs/2511.02824 Kosmos: An AI Scientist for Autonomous Discovery] ([https://edisonscientific.com/ Edison])
 
* 2025-11: [https://arxiv.org/abs/2511.02824 Kosmos: An AI Scientist for Autonomous Discovery] ([https://edisonscientific.com/ Edison])

Latest revision as of 09:40, 2 April 2026

AI Use-cases for Science

Literature

LLM extract data from papers

AI finding links in literature

(Pre) Generate Articles

Explanation

Autonomous Ideation

Adapting LLMs to Science

AI/LLM Control of Scientific Instruments/Facilities

AI/ML Methods tailored to Science

Science Foundation Models

Regression (Data Fitting)

Tabular Classification/Regression

Symbolic Regression

Literature Discovery

Commercial

Bio

AI/ML Methods in Science

Imaging

Materials

Chemistry

Biology

Medicine

See: AI_Agents#Medicine

Successes

AI/ML Methods co-opted for Science

Mechanistic Interpretability

Train large model on science data. Then apply mechanistic interpretability (e.g. sparse autoencoders, SAE) to the feature/activation space.

Uncertainty

Science Benchmarks

Science Agents

Reviews

Challenges

Specific

Science Multi-Agent Setups

Science Agentic Components

Frameworks

Personalities

Skills

AI Science Systems

Inorganic Materials Discovery

Materials Characterization

Chemistry

Bio

Physics

LLMs Optimized for Science

Impact of AI in Science

Related Tools

Literature Search

Data Visualization

Generative

Chemistry

Science Datasets

Genuine Discoveries

Math

Physics assistance

Literature exploration

Bio design

Material Discovery

See Also

Retrieved from "http://gisaxs.com/index.php?title=Science_Agents&oldid=8771"