← 返回论文库

AI Scientist v1: Towards Fully Automated Open-Ended Scientific Discovery

Chris Lu, et al. (Sakana AI) · 2024
L1 · Domain Research AgentsarXiv:2408.06292#research-agent#autonomous
CORE IDEA
第一个端到端自主完成 idea→experiment→paper→review 全流程的系统，含 LLM-based peer review。
CONCRETE EXAMPLE
NanoGPT/2D Diffusion/Grokking 三领域产出 10 篇 paper，平均成本 $15/paper。
L-ANCHOR · 为什么在这一层重要
通用 ML 研究 agent 起点
arXiv:2408.06292 ↗源码 ↗

相关论文

MLR-Copilot: Autonomous Machine Learning Research based on LLM Agents
L0.42024
FunSearch: Mathematical Discoveries from Program Search with LLMs
L12023
AI Scientist v2: Workshop-Level Automated Scientific Discovery via Agentic Tree Search
L12025
ASI-ARCH: AlphaGo Moment for Model Architecture Discovery
L12025