InterLUDE: Interactions between Labeled and Unlabeled Data to Enhance Semi-Supervised Learning
This work addresses the challenge of enhancing semi-supervised learning for image classification, particularly in low-label scenarios, by proposing a novel method that is not incremental but introduces new components for labeled-unlabeled interaction.
The paper tackles the problem of semi-supervised learning by introducing InterLUDE, which leverages interactions between labeled and unlabeled data to improve performance, achieving a 3.2% error rate on STL-10 with 40 labels compared to 14.9% for previous methods.
Semi-supervised learning (SSL) seeks to enhance task performance by training on both labeled and unlabeled data. Mainstream SSL image classification methods mostly optimize a loss that additively combines a supervised classification objective with a regularization term derived solely from unlabeled data. This formulation neglects the potential for interaction between labeled and unlabeled images. In this paper, we introduce InterLUDE, a new approach to enhance SSL made of two parts that each benefit from labeled-unlabeled interaction. The first part, embedding fusion, interpolates between labeled and unlabeled embeddings to improve representation learning. The second part is a new loss, grounded in the principle of consistency regularization, that aims to minimize discrepancies in the model's predictions between labeled versus unlabeled inputs. Experiments on standard closed-set SSL benchmarks and a medical SSL task with an uncurated unlabeled set show clear benefits to our approach. On the STL-10 dataset with only 40 labels, InterLUDE achieves 3.2% error rate, while the best previous method reports 14.9%.