Uncertainty Modeling for Out-of-Distribution Generalization

TL;DR

This paper introduces a probabilistic approach to modeling feature statistics as Gaussian distributions to enhance deep neural networks' robustness and generalization in out-of-distribution scenarios across various vision tasks.

Contribution

It proposes a novel uncertainty modeling method for feature statistics, improving out-of-distribution generalization without adding extra parameters.

Findings

Consistently improves generalization across multiple vision tasks.

Enhances robustness against domain shifts.

Integrates seamlessly into existing networks.

Abstract

Though remarkable progress has been achieved in various vision tasks, deep neural networks still suffer obvious performance degradation when tested in out-of-distribution scenarios. We argue that the feature statistics (mean and standard deviation), which carry the domain characteristics of the training data, can be properly manipulated to improve the generalization ability of deep learning models. Common methods often consider the feature statistics as deterministic values measured from the learned features and do not explicitly consider the uncertain statistics discrepancy caused by potential domain shifts during testing. In this paper, we improve the network generalization ability by modeling the uncertainty of domain shifts with synthesized feature statistics during training. Specifically, we hypothesize that the feature statistic, after considering the potential uncertainties, follows a multivariate Gaussian distribution. Hence, each feature statistic is no longer a deterministic value, but a probabilistic point with diverse distribution possibilities. With the uncertain feature statistics, the models can be trained to alleviate the domain perturbations and achieve better robustness against potential domain shifts. Our method can be readily integrated into networks without additional parameters. Extensive experiments demonstrate that our proposed method consistently improves the network generalization ability on multiple vision tasks, including image classification, semantic segmentation, and instance retrieval. The code can be available at https://github.com/lixiaotong97/DSU.