MaskMatch: Boosting Semi-Supervised Learning Through Mask Autoencoder-Driven Feature Learning

TL;DR

MaskMatch introduces a novel semi-supervised learning algorithm that fully utilizes unlabeled data through Masked Autoencoder-driven feature learning, significantly improving performance on challenging datasets.

Contribution

The paper proposes MaskMatch, a new SSL method integrating Masked Autoencoder for complete data utilization and a synthetic data training approach, advancing beyond threshold-based data filtering.

Findings

Achieves state-of-the-art results on CIFAR-100, STL-10, and Euro-SAT datasets.

Utilizes 100% of available unlabeled data, unlike previous methods.

Demonstrates significant error rate reductions in semi-supervised learning tasks.

Abstract

Conventional methods in semi-supervised learning (SSL) often face challenges related to limited data utilization, mainly due to their reliance on threshold-based techniques for selecting high-confidence unlabeled data during training. Various efforts (e.g., FreeMatch) have been made to enhance data utilization by tweaking the thresholds, yet none have managed to use 100% of the available data. To overcome this limitation and improve SSL performance, we introduce \algo, a novel algorithm that fully utilizes unlabeled data to boost semi-supervised learning. \algo integrates a self-supervised learning strategy, i.e., Masked Autoencoder (MAE), that uses all available data to enforce the visual representation learning. This enables the SSL algorithm to leverage all available data, including samples typically filtered out by traditional methods. In addition, we propose a synthetic data training approach to further increase data utilization and improve generalization. These innovations lead \algo to achieve state-of-the-art results on challenging datasets. For instance, on CIFAR-100 with 2 labels per class, STL-10 with 4 labels per class, and Euro-SAT with 2 labels per class, \algo achieves low error rates of 18.71%, 9.47%, and 3.07%, respectively. The code will be made publicly available.