EvoPool: Evolutionary Programmatic Annotation for Label-Efficient Specialized Supervision

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Xu, Tianyi, Zhang, Yaolun, Ouyang, Xuan, Wang, Huazheng
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916072201912320
author Xu, Tianyi
Zhang, Yaolun
Ouyang, Xuan
Wang, Huazheng
author_facet Xu, Tianyi
Zhang, Yaolun
Ouyang, Xuan
Wang, Huazheng
contents 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
format Preprint
id arxiv_https___arxiv_org_abs_2606_01617
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle EvoPool: Evolutionary Programmatic Annotation for Label-Efficient Specialized Supervision
Xu, Tianyi
Zhang, Yaolun
Ouyang, Xuan
Wang, Huazheng
Computation and Language
Artificial Intelligence
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
title EvoPool: Evolutionary Programmatic Annotation for Label-Efficient Specialized Supervision
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2606.01617