Mitigating Overfitting in Supervised Classification from Two Unlabeled Datasets: A Consistent Risk Correction Approach

Abstract

The recently proposed unlabeled-unlabeled (UU) classification method allows us to train a binary classifier only from two unlabeled datasets with different class priors. Since this method is based on the empirical risk minimization, it works as if it is a supervised classification method, compatible with any model and optimizer. However, this method sometimes suffers from severe overfitting, which we would like to prevent in this paper. Our empirical finding in applying the original UU method is that overfitting often co-occurs with the empirical risk going negative, which is not legitimate. Therefore, we propose to wrap the terms that cause a negative empirical risk by certain correction functions. Then, we prove the consistency of the corrected risk estimator and derive an estimation error bound for the corrected risk minimizer. Experiments show that our proposal can successfully mitigate overfitting of the UU method and significantly improve the classification accuracy.

References in corpus (11)

• Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
• Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithms
• Classification with Noisy Labels by Importance Reweighting
• Distributional Smoothing with Virtual Adversarial Training
• Learning Discrete Representations via Information Maximizing Self-Augmented Training
• Making Risk Minimization Tolerant to Label Noise
• Realistic Evaluation of Deep Semi-Supervised Learning Algorithms
• Estimating the class prior and posterior from noisy positives and unlabeled data
• Theoretical Comparisons of Positive-Unlabeled Learning against Positive-Negative Learning
• On Symmetric Losses for Learning from Corrupted Labels
• On the Minimal Supervision for Training Any Binary Classifier from Only Unlabeled Data

Cited by in corpus (9)

• Do We Need Zero Training Loss After Achieving Zero Training Error?
• Provably Consistent Partial-Label Learning
• Learning with Multiple Complementary Labels
• Rethinking Importance Weighting for Deep Learning under Distribution Shift
• Non-Negative Bregman Divergence Minimization for Deep Direct Density Ratio Estimation
• Pointwise Binary Classification with Pairwise Confidence Comparisons
• Lower-Bounded Proper Losses for Weakly Supervised Classification
• Binary Classification from Multiple Unlabeled Datasets via Surrogate Set Classification
• A Symmetric Loss Perspective of Reliable Machine Learning