EvoPool: Evolutionary Programmatic Annotation for Label-Efficient Specialized Supervision 事件

EvoPool: Evolutionary Programmatic Annotation for Label-Efficient Specialized Supervision arXiv:2606.01617v1 Announce Type: new 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