IDEAgent: Agentic Quality-Diversity Search for Research Idea Generation
IDEAgent frames research ideation as a search for ideas that are both strong and meaningfully different.
The paper introduces a multi-agent system that evolves ideas through lineages, using repair and refinement for quality while checking against completed, rejected, and ancestral ideas for diversity. It proposes Yield, a metric for counting the largest set of mutually diverse ideas above a quality threshold. Across 32 computer science topics, the authors report a 3.89x Yield gain over the best baseline and non-zero Yield on 8x more topics. HF Daily Papers' note
The paper introduces a multi-agent system that evolves ideas through lineages, using repair and refinement for quality while checking against completed, rejected, and ancestral ideas for diversity. It proposes Yield, a metric for counting the largest set of mutually diverse ideas above a quality threshold. Across 32 computer science topics, the authors report a 3.89x Yield gain over the best baseline and non-zero Yield on 8x more topics. HF Daily Papers' note
score 4