Metric-Based Approaches for Semi-Supervised Regression and Classification

Abstract

Semi-supervised learning methods typically require an explicit relationship to be asserted between the labeled and unlabeled data—as illustrated, for example, by the neighbourhoods used in graph-based methods. Semi-supervised model selection and regularization methods are presented here that instead require only that the labeled and unlabeled data are drawn from the same distribution. From this assumption, a metric can be constructed over hypotheses based on their predictions for unlabeled data. This metric can then be used to detect untrustworthy training error estimates, leading to model selection strategies that select the richest hypothesis class while providing theoretical guarantees against over-fitting. This general approach is then adapted to regularization for supervised regression and supervised classification with probabilistic classifiers. The regularization adapts not only to the hypothesis class but also to the specific data sample provided, allowing for better performance than regularizers that account only for class complexity.