Saved in:
Bibliographic Details
Main Authors: Wang, Jiaqi, Zhou, Yuhang, Zhang, Zhixiong, Chen, Qiguang, Chen, Yongqiang, Cheng, James
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2502.12413
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915297300054016
author Wang, Jiaqi
Zhou, Yuhang
Zhang, Zhixiong
Chen, Qiguang
Chen, Yongqiang
Cheng, James
author_facet Wang, Jiaqi
Zhou, Yuhang
Zhang, Zhixiong
Chen, Qiguang
Chen, Yongqiang
Cheng, James
contents Out-of-distribution generalization is a common problem that expects the model to perform well in the different distributions even far from the train data. A popular approach to addressing this issue is invariant learning (IL), in which the model is compiled to focus on invariant features instead of spurious features by adding strong constraints during training. However, there are some potential pitfalls of strong invariant constraints. Due to the limited number of diverse environments and over-regularization in the feature space, it may lead to a loss of important details in the invariant features while alleviating the spurious correlations, namely the over-invariance, which can also degrade the generalization performance. We theoretically define the over-invariance and observe that this issue occurs in various classic IL methods. To alleviate this issue, we propose a simple approach Diverse Invariant Learning (DivIL) by adding the unsupervised contrastive learning and the random masking mechanism compensatory for the invariant constraints, which can be applied to various IL methods. Furthermore, we conduct experiments across multiple modalities across 12 datasets and 6 classic models, verifying our over-invariance insight and the effectiveness of our DivIL framework. Our code is available at https://github.com/kokolerk/DivIL.
format Preprint
id arxiv_https___arxiv_org_abs_2502_12413
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle DivIL: Unveiling and Addressing Over-Invariance for Out-of- Distribution Generalization
Wang, Jiaqi
Zhou, Yuhang
Zhang, Zhixiong
Chen, Qiguang
Chen, Yongqiang
Cheng, James
Machine Learning
Out-of-distribution generalization is a common problem that expects the model to perform well in the different distributions even far from the train data. A popular approach to addressing this issue is invariant learning (IL), in which the model is compiled to focus on invariant features instead of spurious features by adding strong constraints during training. However, there are some potential pitfalls of strong invariant constraints. Due to the limited number of diverse environments and over-regularization in the feature space, it may lead to a loss of important details in the invariant features while alleviating the spurious correlations, namely the over-invariance, which can also degrade the generalization performance. We theoretically define the over-invariance and observe that this issue occurs in various classic IL methods. To alleviate this issue, we propose a simple approach Diverse Invariant Learning (DivIL) by adding the unsupervised contrastive learning and the random masking mechanism compensatory for the invariant constraints, which can be applied to various IL methods. Furthermore, we conduct experiments across multiple modalities across 12 datasets and 6 classic models, verifying our over-invariance insight and the effectiveness of our DivIL framework. Our code is available at https://github.com/kokolerk/DivIL.
title DivIL: Unveiling and Addressing Over-Invariance for Out-of- Distribution Generalization
topic Machine Learning
url https://arxiv.org/abs/2502.12413