FixMatch: Simplifying Semi-Supervised Learning with Consistency and Confidence

TL;DR

FixMatch introduces a simple yet effective semi-supervised learning algorithm that combines consistency regularization and pseudo-labeling, achieving state-of-the-art results on standard benchmarks with minimal labeled data.

Contribution

The paper presents FixMatch, a novel SSL method that simplifies existing approaches by combining consistency and confidence-based pseudo-labeling, leading to improved performance.

Findings

Achieves 94.93% accuracy on CIFAR-10 with 250 labels

Attains 88.61% accuracy with only 4 labels per class

Extensive ablation studies identify key factors for FixMatch's success

Abstract

Semi-supervised learning (SSL) provides an effective means of leveraging unlabeled data to improve a model's performance. In this paper, we demonstrate the power of a simple combination of two common SSL methods: consistency regularization and pseudo-labeling. Our algorithm, FixMatch, first generates pseudo-labels using the model's predictions on weakly-augmented unlabeled images. For a given image, the pseudo-label is only retained if the model produces a high-confidence prediction. The model is then trained to predict the pseudo-label when fed a strongly-augmented version of the same image. Despite its simplicity, we show that FixMatch achieves state-of-the-art performance across a variety of standard semi-supervised learning benchmarks, including 94.93% accuracy on CIFAR-10 with 250 labels and 88.61% accuracy with 40 -- just 4 labels per class. Since FixMatch bears many similarities to existing SSL methods that achieve worse performance, we carry out an extensive ablation study to tease apart the experimental factors that are most important to FixMatch's success. We make our code available at https://github.com/google-research/fixmatch.