SUMix: Mixup with Semantic and Uncertain Information

TL;DR

SUMix introduces a learnable mixing ratio and uncertainty modeling to enhance mixup data augmentation, leading to improved classifier performance across multiple image benchmarks.

Contribution

The paper proposes SUMix, a novel mixup method that learns the mixing ratio and models uncertainty to address semantic information corruption in mixed samples.

Findings

Improves classifier accuracy on five image benchmarks.

Enhances mixup approaches with learned mixing ratios.

Models uncertainty to prevent misleading label information.

Abstract

Mixup data augmentation approaches have been applied for various tasks of deep learning to improve the generalization ability of deep neural networks. Some existing approaches CutMix, SaliencyMix, etc. randomly replace a patch in one image with patches from another to generate the mixed image. Similarly, the corresponding labels are linearly combined by a fixed ratio $\lambda$ by l. The objects in two images may be overlapped during the mixing process, so some semantic information is corrupted in the mixed samples. In this case, the mixed image does not match the mixed label information. Besides, such a label may mislead the deep learning model training, which results in poor performance. To solve this problem, we proposed a novel approach named SUMix to learn the mixing ratio as well as the uncertainty for the mixed samples during the training process. First, we design a learnable similarity function to compute an accurate mix ratio. Second, an approach is investigated as a regularized term to model the uncertainty of the mixed samples. We conduct experiments on five image benchmarks, and extensive experimental results imply that our method is capable of improving the performance of classifiers with different cutting-based mixup approaches. The source code is available at https://github.com/JinXins/SUMix.