Classification from Pairwise Similarity and Unlabeled Data
arXiv:1802.04381
Abstract
Supervised learning needs a huge amount of labeled data, which can be a big bottleneck under the situation where there is a privacy concern or labeling cost is high. To overcome this problem, we propose a new weakly-supervised learning setting where only similar (S) data pairs (two examples belong to the same class) and unlabeled (U) data points are needed instead of fully labeled data, which is called SU classification. We show that an unbiased estimator of the classification risk can be obtained only from SU data, and the estimation error of its empirical risk minimizer achieves the optimal parametric convergence rate. Finally, we demonstrate the effectiveness of the proposed method through experiments.
References in corpus (7)
- Learning Discrete Representations via Information Maximizing Self-Augmented Training
- Positive-Unlabeled Learning with Non-Negative Risk Estimator
- Prediction and Clustering in Signed Networks: A Local to Global Perspective
- Clustering Partially Observed Graphs via Convex Optimization
- Mixture Proportion Estimation via Kernel Embedding of Distributions
- Theoretical Comparisons of Positive-Unlabeled Learning against Positive-Negative Learning
- Information-theoretic Semi-supervised Metric Learning via Entropy Regularization