Reproduce “DualEvolve: Dual-Level Self-Evolution of Mechanism and Orchestration for Recommendation” 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

Tianqi Gao, Chenxu Wang, Wenxuan Yang, Yueqi Mou, Biao Tang, Ming Liu, Peng Yan

DualEvolve co-evolves recommendation code and the orchestration rules that guide experiments, while keeping evaluation and promotion criteria fixed. Its evaluation covers four datasets and four models, ranking one held-out item against 99 negatives. It exceeds three baselines in all 64 setting–metric comparisons, with macro-averaged gains over DeepEvolve of 7.6% on HR@5 and 9.0% on NDCG@5; these are not online results.

Summary by alphaXiv.

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