Analysis of Kernel Mean Matching under Covariate Shift

Abstract

In real supervised learning scenarios, it is not uncommon that the training and test sample follow different probability distributions, thus rendering the necessity to correct the sampling bias. Focusing on a particular covariate shift problem, we derive high probability confidence bounds for the kernel mean matching (KMM) estimator, whose convergence rate turns out to depend on some regularity measure of the regression function and also on some capacity measure of the kernel. By comparing KMM with the natural plug-in estimator, we establish the superiority of the former hence provide concrete evidence/understanding to the effectiveness of KMM under covariate shift.

ICML2012

Cited by in corpus (15)

• A review of domain adaptation without target labels
• An introduction to domain adaptation and transfer learning
• Domain Generalization by Marginal Transfer Learning
• Domain Adaptation with Asymmetrically-Relaxed Distribution Alignment
• A Review of Single-Source Deep Unsupervised Visual Domain Adaptation
• Characterizing and Avoiding Negative Transfer
• On Minimax Optimal Offline Policy Evaluation
• Open Set Domain Adaptation: Theoretical Bound and Algorithm
• Normalized Wasserstein Distance for Mixture Distributions with Applications in Adversarial Learning and Domain Adaptation
• A Generalized Neyman-Pearson Criterion for Optimal Domain Adaptation
• Transfer String Kernel for Cross-Context DNA-Protein Binding Prediction
• Robust Importance Weighting for Covariate Shift
• Structured Discriminative Tensor Dictionary Learning for Unsupervised Domain Adaptation
• Doubly Robust Stein-Kernelized Monte Carlo Estimator: Simultaneous Bias-Variance Reduction and Supercanonical Convergence
• Discriminative Density-ratio Estimation