Diffusion Models and Semi-Supervised Learners Benefit Mutually with Few Labels

Published: 21 Sept 2023, Last Modified: 02 Nov 2023NeurIPS 2023 spotlightEveryoneRevisionsBibTeX
Keywords: diffusion models, semi-supervised generation, semi-supervised diffusion models, semi-supervised classification, image generation.
TL;DR: Diffusion models and semi-supervised learners can benefit mutually to achieve SOTA semi-supervised classification and generation results on ImageNet-1K and CIFAR-10 with few labels.
Abstract: In an effort to further advance semi-supervised generative and classification tasks, we propose a simple yet effective training strategy called *dual pseudo training* (DPT), built upon strong semi-supervised learners and diffusion models. DPT operates in three stages: training a classifier on partially labeled data to predict pseudo-labels; training a conditional generative model using these pseudo-labels to generate pseudo images; and retraining the classifier with a mix of real and pseudo images. Empirically, DPT consistently achieves SOTA performance of semi-supervised generation and classification across various settings. In particular, with one or two labels per class, DPT achieves a Fréchet Inception Distance (FID) score of 3.08 or 2.52 on ImageNet $256\times256$. Besides, DPT outperforms competitive semi-supervised baselines substantially on ImageNet classification tasks, *achieving top-1 accuracies of 59.0 (+2.8), 69.5 (+3.0), and 74.4 (+2.0)* with one, two, or five labels per class, respectively. Notably, our results demonstrate that diffusion can generate realistic images with only a few labels (e.g., $<0.1$%) and generative augmentation remains viable for semi-supervised classification. Our code is available at *https://github.com/ML-GSAI/DPT*.
Supplementary Material: zip
Submission Number: 6170