Collaborative evidence sharing speeds up ai scientific discovery
Collaborative Principle Evolution via Evidence Transfer for Scientific Discovery
Machine Learning
Summary
Figuring out new scientific ideas using AI can take a long time because there are so many possibilities to test. The authors created a system called COEVOLVE that lets multiple AI processes work together by sharing helpful information while still exploring different ideas on their own. This teamwork lets COEVOLVE find better solutions faster than previous methods that worked alone. They tested it on various scientific tasks and found it both more accurate and quicker, showing when sharing information helps and when it's best to be cautious.
What this means in practice
- •For autonomous research teams: Improve automated scientific research pipelines by enabling parallel hypothesis exploration with information sharing to find solutions faster and better.
- •For ai development teams: Enhance large language model-based AI systems with coordinated evidence sharing across parallel reasoning branches for more efficient problem solving.
Authors
Yingming Pu, Hongyu Chen, Tao Lin
Abstract
Large Language Model (LLM)-based agents promise to automate scientific discovery, yet exploring the vast hypothesis space remains costly. Existing principle-evolution methods accelerate this loop, but operate sequentially, which caps exploration breadth and wastes wall-clock time on challenging problems. To address this, we formulate collaborative scientific discovery as evidence transfer between parallel principle-evolution branches. We present COEVOLVE, which realizes this transfer through a coordination core over parallel branches. By integrating value-of-information-gated routing and context-discounted likelihood injection, COEVOLVE enables branches to collaborate through shared measurements while keeping their principle posteriors separate. Across six scientific-discovery tasks under a matched evaluation budget, COEVOLVE attains a mean solution quality of 66.5% versus 57.0% for single-branch principle evolution, with a 1.80x mean wall-clock speedup on the GPT-5.6-Terra backbone; on five auto-research tasks delegated to an autonomous research harness, it is the only arm whose mean stays above the published SOTA anchor on every task. These results establish when evidence sharing accelerates parallel discovery and when transfer safeguards are necessary to limit negative or inert transfers