The AI Scientist-v2: Workshop-Level Automated Scientific Discovery via Agentic Tree Search
Summary
AI Scientist-v2 is a generalized end-to-end agentic system for autonomous scientific discovery, capable of generating hypotheses, conducting experiments, analyzing data, and drafting manuscripts. It advances beyond the v1 version by removing templates and using progressive agentic tree search to explore across ML domains, with sandboxing and cost-awareness emphasized. The project includes installation steps, ideation and experiment pipelines, and references to its workshop paper, blog, and ICLR2025 workshop materials.