Accelerating scientific discovery with Co-Scientist - PubMed
5 hours ago
- #AI in science
- #Hypothesis generation
- #Drug repurposing
- Co-Scientist is a multi-agent AI system using Gemini for structured scientific thinking and hypothesis generation.
- It employs agents to continuously generate, critique, and refine hypotheses, enhanced by scaling test-time compute.
- Key features include a multi-agent architecture with asynchronous task execution and a tournament evolution process for self-improvement.
- Automated evaluations show that scaling test-time compute improves hypothesis quality over time.
- Validated in biomedical applications: drug repurposing, novel target discovery, and explaining antimicrobial resistance mechanisms.
- Specifically, it identified new drug repurposing candidates and synergistic combination therapies for acute myeloid leukemia, validated in vitro.