Mixup Regularization: A Probabilistic Perspective

TL;DR

This paper introduces a probabilistic framework for mixup regularization tailored for conditional density estimation, enabling analytical likelihood fusion and improved model generalization.

Contribution

It presents a novel probabilistic mixup approach based on likelihood fusion for exponential family distributions, extending mixup to intermediate neural network layers.

Findings

Analytical likelihood fusion improves regularization effectiveness.

Extending mixup to intermediate layers enhances model performance.

Empirical results outperform existing mixup variants on synthetic and real data.

Abstract

In recent years, mixup regularization has gained popularity as an effective way to improve the generalization performance of deep learning models by training on convex combinations of training data. While many mixup variants have been explored, the proper adoption of the technique to conditional density estimation and probabilistic machine learning remains relatively unexplored. This work introduces a novel framework for mixup regularization based on probabilistic fusion that is better suited for conditional density estimation tasks. For data distributed according to a member of the exponential family, we show that likelihood functions can be analytically fused using log-linear pooling. We further propose an extension of probabilistic mixup, which allows for fusion of inputs at an arbitrary intermediate layer of the neural network. We provide a theoretical analysis comparing our approach to standard mixup variants. Empirical results on synthetic and real datasets demonstrate the benefits of our proposed framework compared to existing mixup variants.