Domain Adaptation under Open Set Label Shift
Domain Adaptation under Open Set Label Shift
复制标题
DOI：
10.48550/arxiv.2207.13048
复制
发表时间：
2022-07
期刊：
影响因子：
--
通讯作者：
S. Garg;Sivaraman Balakrishnan;Zachary Chase Lipton
中科院分区：
文献类型：
--
作者：
S. Garg;Sivaraman Balakrishnan;Zachary Chase Lipton
我们引入了开放集标签移位（Open Set Label Shift， OSLS）下的领域自适应问题，其中标签分布可以任意改变，并且在部署过程中可能会出现新的类，但类条件分布p（x|y）是领域不变的。ols包含了标签移位和正无标签学习下的领域自适应。这里学习者的目标有两个：(a)估计目标标签分布，包括新类；(b)学习目标分类器。首先，我们建立了确定这些量的充分必要条件。其次，在标签转移和PU学习进展的推动下，我们为这两个任务提出了利用黑箱预测器的实用方法。与典型的开放集域自适应（OSDA）问题不同，OSDA问题往往是病态的，只适用于启发式，而OSLS提供了一个适合于更多原则性机制的良好定位问题。在视觉、语言和医疗数据集的许多半合成基准上进行的实验表明，我们的方法始终优于OSDA基线，在目标域精度上实现了10- 25%的提高。最后，我们分析了所提出的方法，建立了在高斯设置下线性模型的有限样本收敛到真标签边缘和收敛到最优分类器。代码可从https://github.com/acmi-lab/Open-Set-Label-Shift获得。
We introduce the problem of domain adaptation under Open Set Label Shift (OSLS) where the label distribution can change arbitrarily and a new class may arrive during deployment, but the class-conditional distributions p(x|y) are domain-invariant. OSLS subsumes domain adaptation under label shift and Positive-Unlabeled (PU) learning. The learner's goals here are two-fold: (a) estimate the target label distribution, including the novel class; and (b) learn a target classifier. First, we establish necessary and sufficient conditions for identifying these quantities. Second, motivated by advances in label shift and PU learning, we propose practical methods for both tasks that leverage black-box predictors. Unlike typical Open Set Domain Adaptation (OSDA) problems, which tend to be ill-posed and amenable only to heuristics, OSLS offers a well-posed problem amenable to more principled machinery. Experiments across numerous semi-synthetic benchmarks on vision, language, and medical datasets demonstrate that our methods consistently outperform OSDA baselines, achieving 10--25% improvements in target domain accuracy. Finally, we analyze the proposed methods, establishing finite-sample convergence to the true label marginal and convergence to optimal classifier for linear models in a Gaussian setup. Code is available at https://github.com/acmi-lab/Open-Set-Label-Shift.