RankEvolve: Automating the Discovery of Retrieval Algorithms via LLM-Driven Evolution

TL;DR

This paper presents RankEvolve, a method that uses large language models and evolutionary algorithms to automatically discover improved retrieval algorithms, starting from traditional methods like BM25, showing promising results across multiple IR datasets.

Contribution

Introducing RankEvolve, a novel framework that automates the discovery of retrieval algorithms using LLM-guided evolutionary search, starting from standard IR methods.

Findings

Evolved algorithms outperform baseline methods on multiple datasets.

Evolved algorithms show good transferability to unseen benchmarks.

Evaluator-guided LLM evolution is effective for algorithm discovery.

Abstract

Retrieval algorithms like BM25 and query likelihood with Dirichlet smoothing remain strong and efficient first-stage rankers, yet improvements have mostly relied on parameter tuning and human intuition. We investigate whether a large language model, guided by an evaluator and evolutionary search, can automatically discover improved lexical retrieval algorithms. We introduce RankEvolve, a program evolution setup based on AlphaEvolve, in which candidate ranking algorithms are represented as executable code and iteratively mutated, recombined, and selected based on retrieval performance across 12 IR datasets from BEIR and BRIGHT. RankEvolve starts from two seed programs: BM25 and query likelihood with Dirichlet smoothing. The evolved algorithms are novel, effective, and show promising transfer to the full BEIR and BRIGHT benchmarks as well as TREC DL 19 and 20. Our results suggest that evaluator-guided LLM program evolution is a practical path towards automatic discovery of novel ranking algorithms.