IDEAgent frames research ideation as a Quality-Diversity (QD) search problem rather than optimizing quality or diversity separately. Its multi-agent framework evolves ideas through lineages, using multi-objective feedback for repair and refinement, plus sequential memory and explicit comparisons with completed ideas, ancestors, and rejected proposals. The paper introduces Yield, a metric measuring the largest mutually diverse set of ideas above a quality threshold. Across 32 topics spanning 8 computer science domains, IDEAgent reportedly achieves 3.89x the Yield of the strongest baseline and non-zero Yield on eight times more topics. The implementation is open-sourced.
No heat snapshots are available in the last 24 hours.