Adaptive Discriminative Regularization for Visual Classification

TL;DR

This paper introduces Adaptive Discriminative Regularization (ADR), a novel method that enhances discriminative feature learning by accounting for semantic overlaps among classes, improving classification performance especially in complex, real-world data distributions.

Contribution

The paper proposes ADR, a new regularization technique that leverages class semantic overlaps and adaptive penalties to improve discriminative learning in visual classification tasks.

Findings

Consistent performance improvements across over 10 benchmarks.

Robustness to long-tailed and noisy label distributions.

Compatible with mainstream architectures and loss functions.

Abstract

How to improve discriminative feature learning is central in classification. Existing works address this problem by explicitly increasing inter-class separability and intra-class similarity, whether by constructing positive and negative pairs for contrastive learning or posing tighter class separating margins. These methods do not exploit the similarity between different classes as they adhere to i.i.d. assumption in data. In this paper, we embrace the real-world data distribution setting that some classes share semantic overlaps due to their similar appearances or concepts. Regarding this hypothesis, we propose a novel regularization to improve discriminative learning. We first calibrate the estimated highest likelihood of one sample based on its semantically neighboring classes, then encourage the overall likelihood predictions to be deterministic by imposing an adaptive exponential penalty. As the gradient of the proposed method is roughly proportional to the uncertainty of the predicted likelihoods, we name it adaptive discriminative regularization (ADR), trained along with a standard cross entropy loss in classification. Extensive experiments demonstrate that it can yield consistent and non-trivial performance improvements in a variety of visual classification tasks (over 10 benchmarks). Furthermore, we find it is robust to long-tailed and noisy label data distribution. Its flexible design enables its compatibility with mainstream classification architectures and losses.