Reproduce “IDEAgent: Agentic Quality-Diversity Search for Research Idea Generation” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Varun Gumma, Navonil Majumder, Soumitra Sinhahajari, Soujanya Poria

IDEAgent searches for portfolios of distinct ideas by generating proposals, recording summaries of active, historical, and rejected ideas, and using critique to repair or refine candidates. Yield counts the largest subset passing quality thresholds whose every pair meets a distinctness threshold. Across 32 topics, mean Yield was 1.094 versus NOVA’s 0.281 under LLM evaluation; not literature novelty or validation.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.