Denoising Mixup for Regression

Hou, Zhengzhang and Li, Zhanshan and Liu, Yanbo and Nitschke, Geoff and Lu, You and Li, Ximing (2026) Denoising Mixup for Regression, Proceedings of 40th Annual AAAI Conference on Artificial Intelligence, 20-27 January 2026, Singapore, 21805-21813, AAAI Press.

Text 2026-denoising-mixup-for-regression.pdf Download (468kB)

Abstract

Data augmentation is an intuitive solution to increase the diversity of training instances in the machine learning community. Mixup is acknowledged as an effective and efficient mixbased data augmentation method, following a linear alignment assumption that the linear interpolations of features align the corresponding linear interpolations of labels. Unfortunately, this assumption can be violated in many complex scenarios, resulting in augmented instances with noisy labels, especially for regression problems. To solve this problem, we propose an easy-to-implement mixup method, namely DEnosing MIXUP (DE-MIXUP), which iteratively corrects the noisy response targets by leveraging an auxiliary noise estimation task with mixup deep features. Additionally, we suggest an efficient optimization method with alternating direction method of multipliers. We compare DE-MIXUP with the existing mixup variants and other prevalent data augmentation methods across benchmark regression datasets. Empirical results indicate the effectiveness of DE-MIXUP under the in-distribution and out-of-distribution cases.

Item Type:	Conference paper
Subjects:	Computing methodologies > Machine learning
Date Deposited:	15 Oct 2026 06:59
Last Modified:	15 Oct 2026 06:59
URI:	https://pubs.cs.uct.ac.za/id/eprint/1819

Actions (login required)

View Item