Mixture Proportion Estimation via Kernel Embedding of Distributions

Abstract

Mixture proportion estimation (MPE) is the problem of estimating the weight of a component distribution in a mixture, given samples from the mixture and component. This problem constitutes a key part in many "weakly supervised learning" problems like learning with positive and unlabelled samples, learning with label noise, anomaly detection and crowdsourcing. While there have been several methods proposed to solve this problem, to the best of our knowledge no efficient algorithm with a proven convergence rate towards the true proportion exists for this problem. We fill this gap by constructing a provably correct algorithm for MPE, and derive convergence rates under certain assumptions on the distribution. Our method is based on embedding distributions onto an RKHS, and implementing it only requires solving a simple convex quadratic programming problem a few times. We run our algorithm on several standard classification datasets, and demonstrate that it performs comparably to or better than other algorithms on most datasets.

References in corpus (2)

• Learning Topic Models - Going beyond SVD
• Nonparametric semi-supervised learning of class proportions

Cited by in corpus (28)

• Learning from positive and unlabeled data: a survey
• Are Anchor Points Really Indispensable in Label-Noise Learning?
• Positive-Unlabeled Learning with Non-Negative Risk Estimator
• Dual T: Reducing Estimation Error for Transition Matrix in Label-noise Learning
• Theoretical Comparisons of Positive-Unlabeled Learning against Positive-Negative Learning
• Semi-Supervised Classification Based on Classification from Positive and Unlabeled Data
• L_DMI: An Information-theoretic Noise-robust Loss Function
• Noise-tolerant fair classification
• Transfer Learning with Label Noise
• Learning with Bounded Instance- and Label-dependent Label Noise
• Learning Noise Transition Matrix from Only Noisy Labels via Total Variation Regularization
• Learning from Positive and Unlabeled Data with Arbitrary Positive Shift
• Decontamination of Mutual Contamination Models
• On the Resistance of Nearest Neighbor to Random Noisy Labels
• Classification from Pairwise Similarity and Unlabeled Data
• A Semi-Supervised Two-Stage Approach to Learning from Noisy Labels
• Classification from Positive, Unlabeled and Biased Negative Data
• Temporal Positive-unlabeled Learning for Biomedical Hypothesis Generation via Risk Estimation
• Partial Optimal Transport with Applications on Positive-Unlabeled Learning
• A Unified View of Label Shift Estimation
• An Unbiased Risk Estimator for Learning with Augmented Classes
• Principled analytic classifier for positive-unlabeled learning via weighted integral probability metric
• When Optimizing $f$-divergence is Robust with Label Noise
• Information-Theoretic Representation Learning for Positive-Unlabeled Classification
• Limited Gradient Descent: Learning With Noisy Labels
• Learning Classifiers on Positive and Unlabeled Data with Policy Gradient
• Positive-Unlabeled Classification under Class Prior Shift and Asymmetric Error
• Identifying Different Definitions of Future in the Assessment of Future Economic Conditions: Application of PU Learning and Text Mining