{"ID":2921229,"CreatedAt":"2026-06-02T02:42:49.606572591Z","UpdatedAt":"2026-06-04T00:54:56.190393508Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2606.01617","arxiv_id":"2606.01617","title":"EvoPool: Evolutionary Programmatic Annotation for Label-Efficient Specialized Supervision","abstract":"Large language models excel at general tasks but underperform smaller supervised models in specialized, high-stakes domains where training labels are costly. We address this regime with EvoPool, an evolutionary multi-agent framework inspired by Darwinian evolution. Three specialized agents iteratively propose executable annotator code, a small validation set provides a fitness signal, and a deterministic gate keeps only annotators that pass viability, diversity, and marginal-contribution checks across generations. Pool votes are mapped to soft training labels by EvoAgg, a text-aware aggregator combining semantic features with annotator-vote features. The authored pool runs at near-zero per-example cost and is 4500 to 31000x faster than LLM annotation on 100K examples. Across 7 of 8 LLM-weak specialized and complex tasks spanning biomedical relation extraction, legal-clause classification, complex reasoning, and dense multi-label biomedical classification, EvoPool beats the strongest LLM annotation baseline by an average +0.141 macro-F1, peaking at +0.301 on ChemProt and +0.265 on PubMed. Code is available at: https://github.com/tianyi0216/EvoPool","short_abstract":"Large language models excel at general tasks but underperform smaller supervised models in specialized, high-stakes domains where training labels are costly. We address this regime with EvoPool, an evolutionary multi-agent framework inspired by Darwinian evolution. Three specialized agents iteratively propose executabl...","url_abs":"https://arxiv.org/abs/2606.01617","url_pdf":"https://arxiv.org/pdf/2606.01617v1","authors":"[\"Tianyi Xu\",\"Yaolun Zhang\",\"Xuan Ouyang\",\"Huazheng Wang\"]","published":"2026-06-01T03:10:37Z","proceeding":"cs.CL","tasks":"[\"cs.CL\",\"cs.AI\"]","methods":"[\"Large Language Model\",\"Language Model\"]","has_code":false,"code_links":[{"ID":612575,"CreatedAt":"2026-06-02T02:42:49.606572591Z","UpdatedAt":"2026-06-02T02:42:49.606572591Z","DeletedAt":null,"paper_id":2921229,"paper_url":"https://arxiv.org/abs/2606.01617","paper_title":"EvoPool: Evolutionary Programmatic Annotation for Label-Efficient Specialized Supervision","repo_url":"https://github.com/tianyi0216/EvoPool","is_official":false,"mentioned_in_paper":false,"mentioned_in_github":true,"github_stars":0}]}