ELSA: Efficient Label Shift Adaptation through the Lens of Semiparametric Models
This addresses label shift adaptation for machine learning practitioners by offering a computationally efficient solution, though it appears incremental as it builds on existing frameworks.
The paper tackles the domain adaptation problem with label shift, where the label distribution changes between training and testing, by proposing ELSA, an efficient method that avoids post-prediction calibrations and achieves state-of-the-art estimation performance.
We study the domain adaptation problem with label shift in this work. Under the label shift context, the marginal distribution of the label varies across the training and testing datasets, while the conditional distribution of features given the label is the same. Traditional label shift adaptation methods either suffer from large estimation errors or require cumbersome post-prediction calibrations. To address these issues, we first propose a moment-matching framework for adapting the label shift based on the geometry of the influence function. Under such a framework, we propose a novel method named \underline{E}fficient \underline{L}abel \underline{S}hift \underline{A}daptation (ELSA), in which the adaptation weights can be estimated by solving linear systems. Theoretically, the ELSA estimator is $\sqrt{n}$-consistent ($n$ is the sample size of the source data) and asymptotically normal. Empirically, we show that ELSA can achieve state-of-the-art estimation performances without post-prediction calibrations, thus, gaining computational efficiency.