Scaling Discovery Requires Forgetting in Automated Research
This paper investigates scaling automated research with LLM agents through iterative refinement, parallel execution, and communication.
This paper investigates scaling automated research with LLM agents through iterative refinement, parallel execution, and communication.