[论文] BatchBench: Benchmarking AI Robustness to Batch Effects and Domain Shi...

论文概要

研究领域: ML 作者: Sonja Topf, Ana Sanchez-Fernandez, Gunter Klambauer 发布时间: 2026-10-21 arXiv: 2610.30104

中文摘要

尽管近期取得了进展，当前的 AI 系统仍然极易受到域偏移的影响，尤其是在生物医学成像领域——图像通常在与训练数据截然不同的实验条件下获取。虽然最近的域适应方法取得了实质性进展，但在域偏移下的显著性能差距仍然存在。为衡量这方面的进展，我们引入 BatchBench——一个评估和比较 AI 系统在来自未见过实验批次和数据源的生物医学图像上鲁棒性的排行榜。BatchBench 基于 JUMP-CP 构建，这是最大的公开可用生物医学成像数据集之一。JUMP-CP 使用 Cell Painting 测定法生成，表现出显著的批次效应和域偏移，使其成为评估分布变化鲁棒性的理想试验场。

原文摘要

Despite recent advances, current AI systems remain highly susceptible to domain shifts, particularly in biomedical imaging, where images are usually acquired under substantially different experimental conditions. Although recent domain adaptation methods have made substantial progress, significant performance gaps under domain shift persist. To measure progress in this regard, we introduce BatchBench, a leaderboard that evaluates and compares the robustness of AI systems on biomedical images from unseen experimental batches and data sources. BatchBench is built on JUMP-CP, one of the largest publicly available biomedical imaging datasets. Generated using the Cell Painting assay, JUMP-CP exhibits pronounced batch effects and domain shifts, making it a challenging testbed for evaluating robu...

#论文 #arXiv #ML #小凯