Preserving Fine-Grain Feature Information in Classification via Entropic Regularization

TL;DR

This paper introduces an entropy-based regularization method to improve fine-grain classification performance when only coarse labels are available, addressing feature information loss caused by standard cross-entropy training.

Contribution

The paper proposes a novel entropy regularization technique that enhances feature diversity, enabling models to better capture fine-grain details from coarse-labeled data.

Findings

Regularization improves fine-grain classification accuracy.

Theoretical analysis supports the effectiveness of entropy regularization.

Empirical results show better feature preservation and model performance.

Abstract

Labeling a classification dataset implies to define classes and associated coarse labels, that may approximate a smoother and more complicated ground truth. For example, natural images may contain multiple objects, only one of which is labeled in many vision datasets, or classes may result from the discretization of a regression problem. Using cross-entropy to train classification models on such coarse labels is likely to roughly cut through the feature space, potentially disregarding the most meaningful such features, in particular losing information on the underlying fine-grain task. In this paper we are interested in the problem of solving fine-grain classification or regression, using a model trained on coarse-grain labels only. We show that standard cross-entropy can lead to overfitting to coarse-related features. We introduce an entropy-based regularization to promote more diversity in the feature space of trained models, and empirically demonstrate the efficacy of this methodology to reach better performance on the fine-grain problems. Our results are supported through theoretical developments and empirical validation.