MLAILGMEOct 2, 2020

Effective Sample Size, Dimensionality, and Generalization in Covariate Shift Adaptation

arXiv:2010.01184v521 citations
Originality Incremental advance
AI Analysis

This provides a unified theoretical framework for domain adaptation in supervised learning, addressing a known bottleneck but is incremental in nature.

The paper tackles the problem of understanding how effective sample size, dimensionality, and generalization are related in covariate shift adaptation, showing that dimensionality reduction can increase effective sample size and improve performance.

In supervised learning, training and test datasets are often sampled from distinct distributions. Domain adaptation techniques are thus required. Covariate shift adaptation yields good generalization performance when domains differ only by the marginal distribution of features. Covariate shift adaptation is usually implemented using importance weighting, which may fail, according to common wisdom, due to small effective sample sizes (ESS). Previous research argues this scenario is more common in high-dimensional settings. However, how effective sample size, dimensionality, and model performance/generalization are formally related in supervised learning, considering the context of covariate shift adaptation, is still somewhat obscure in the literature. Thus, a main challenge is presenting a unified theory connecting those points. Hence, in this paper, we focus on building a unified view connecting the ESS, data dimensionality, and generalization in the context of covariate shift adaptation. Moreover, we also demonstrate how dimensionality reduction or feature selection can increase the ESS, and argue that our results support dimensionality reduction before covariate shift adaptation as a good practice.

Code Implementations1 repo
Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes