Abstract:To address the issues of MixUp and its variants (e.g., Manifold MixUp) in image classification tasks-namely, their neglect of mixing within the same class (intra-class mixup) and their inadequacy in enhancing intra-class cohesion through their mixing operations-we propose a novel mixup method named SynerMix-Intra and, building upon this, introduce a synergistic mixup solution named SynerMix. SynerMix-Intra specifically targets intra-class mixup to bolster intra-class cohesion, a feature not addressed by current mixup methods. For each mini-batch, it leverages feature representations of unaugmented original images from each class to generate a synthesized feature representation through random linear interpolation. All synthesized representations are then fed into the classification and loss layers to calculate an average classification loss that significantly enhances intra-class cohesion. Furthermore, SynerMix combines SynerMix-Intra with an existing mixup approach (e.g., MixUp, Manifold MixUp), which primarily focuses on inter-class mixup and has the benefit of enhancing inter-class separability. In doing so, it integrates both inter- and intra-class mixup in a balanced way while concurrently improving intra-class cohesion and inter-class separability. Experimental results on six datasets show that SynerMix achieves a 0.1% to 3.43% higher accuracy than the best of either MixUp or SynerMix-Intra alone, averaging a 1.16% gain. It also surpasses the top-performer of either Manifold MixUp or SynerMix-Intra by 0.12% to 5.16%, with an average gain of 1.11%. Given that SynerMix is model-agnostic, it holds significant potential for application in other domains where mixup methods have shown promise, such as speech and text classification. Our code is publicly available at:

ISM: intra-class similarity mixing for time series augmentation

Infinite Class Mixup

Intra-Class Adaptive Augmentation with Neighbor Correction for Deep Metric Learning

DoubleMix: Simple Interpolation-Based Data Augmentation for Text Classification

Embarrassingly Simple MixUp for Time-series

Decoupled Mixup for Data-efficient Learning

MixIR: Mixing Input and Representations for Contrastive Learning

MiAMix: Enhancing Image Classification through a Multi-stage Augmented Mixed Sample Data Augmentation Method

MSMix:An Interpolation-Based Text Data Augmentation Method Manifold Swap Mixup

Harnessing Hard Mixed Samples with Decoupled Regularizer

SynerMix: Synergistic Mixup Solution for Enhanced Intra-Class Cohesion and Inter-Class Separability in Image Classification

TransformMix: Learning Transformation and Mixing Strategies from Data

Tailoring Mixup to Data for Calibration

Global Mixup: Eliminating Ambiguity with Clustering.

C-Mixup: Improving Generalization in Regression

ShuffleMix: Improving Representations via Channel-Wise Shuffle of Interpolated Hidden States

Mixup Augmentation with Multiple Interpolations

SnapMix: Semantically Proportional Mixing for Augmenting Fine-grained Data

Mixing Approach for Text Data Augmentation Based on an Ensemble of Explainable Artificial Intelligence Methods

Asymptotic Midpoint Mixup for Margin Balancing and Moderate Broadening

WeMix: How to Better Utilize Data Augmentation