GenLabel: Mixup Relabeling using Generative Models

01/07/2022
by   Jy-yong Sohn, et al.
0

Mixup is a data augmentation method that generates new data points by mixing a pair of input data. While mixup generally improves the prediction performance, it sometimes degrades the performance. In this paper, we first identify the main causes of this phenomenon by theoretically and empirically analyzing the mixup algorithm. To resolve this, we propose GenLabel, a simple yet effective relabeling algorithm designed for mixup. In particular, GenLabel helps the mixup algorithm correctly label mixup samples by learning the class-conditional data distribution using generative models. Via extensive theoretical and empirical analysis, we show that mixup, when used together with GenLabel, can effectively resolve the aforementioned phenomenon, improving the generalization performance and the adversarial robustness.

READ FULL TEXT

page 9

page 39

research
07/23/2023

Improving Out-of-Distribution Robustness of Classifiers via Generative Interpolation

Deep neural networks achieve superior performance for learning from inde...
research
09/07/2019

On Need for Topology Awareness of Generative Models

Manifold assumption in learning states that: the data lie approximately ...
research
07/26/2023

Regularizing Neural Networks with Meta-Learning Generative Models

This paper investigates methods for improving generative data augmentati...
research
05/26/2019

Classification Accuracy Score for Conditional Generative Models

Deep generative models (DGMs) of images are now sufficiently mature that...
research
04/03/2023

A Guide for Practical Use of ADMG Causal Data Augmentation

Data augmentation is essential when applying Machine Learning in small-d...
research
04/13/2020

Data augmentation using generative networks to identify dementia

Data limitation is one of the most common issues in training machine lea...
research
12/01/2022

Why Are Conditional Generative Models Better Than Unconditional Ones?

Extensive empirical evidence demonstrates that conditional generative mo...

Please sign up or login with your details

Forgot password? Click here to reset