arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1903.09239stat.MLcs.LG
收藏

多领域对抗学习

Multi-Domain Adversarial Learning

Department of Pharmaceutical ChemistryUCSF(加州大学旧金山分校药学化学系);INRIA-CNRS-UPSud-UPSaclayTAU, U. Paris-Sud(巴黎-萨克雷大学(U. Paris-Sud))

Alice Schoenauer-Sebag, Louise Heinrich, Marc Schoenauer, Michele Sebag, Lani F. Wu, Steve J. Altschuler

摘要:

多域学习(MDL)的目标是获得一个跨多个域的平均风险最小的模型。我们的经验动机是自动显微镜数据,其中培养的细胞在暴露于已知和未知的化学扰动后成像,并且每个数据集都显示出显着的实验偏差。本文提出了一种多域对抗学习方法MuLANN,用于在半监督环境中利用具有重叠但不同类集的多个数据集。我们的贡献包括:i)使用H-发散获得的MDL中平均和最差域风险的界限; ii)适应半监督多域学习和域适应的新损失; iii)该方法的实验验证,改进了两个标准图像基准和一个新的生物图像数据集Cell的最新技术水平。

英文摘要:

Multi-domain learning (MDL) aims at obtaining a model with minimal average risk across multiple domains. Our empirical motivation is automated microscopy data, where cultured cells are imaged after being exposed to known and unknown chemical perturbations, and each dataset displays significant experimental bias. This paper presents a multi-domain adversarial learning approach, MuLANN, to leverage multiple datasets with overlapping but distinct class sets, in a semi-supervised setting. Our contributions include: i) a bound on the average- and worst-domain risk in MDL, obtained using the H-divergence; ii) a new loss to accommodate semi-supervised multi-domain learning and domain adaptation; iii) the experimental validation of the approach, improving on the state of the art on two standard image benchmarks, and a novel bioimage dataset, Cell.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑