Feature Extraction Framework based on Contrastive Learning with Adaptive Positive and Negative Samples

TL;DR

This paper introduces CL-FEFA, a contrastive learning-based feature extraction framework that adaptively constructs positive and negative samples, improving robustness and discriminability across various supervision settings.

Contribution

The framework adaptively constructs positive and negative samples using feature results and potential structure information, enhancing feature extraction accuracy and robustness.

Findings

Outperforms traditional feature extraction methods

Effective in unsupervised, supervised, and semi-supervised settings

Improves intra-class compactness and inter-class dispersion

Abstract

In this study, we propose a feature extraction framework based on contrastive learning with adaptive positive and negative samples (CL-FEFA) that is suitable for unsupervised, supervised, and semi-supervised single-view feature extraction. CL-FEFA constructs adaptively the positive and negative samples from the results of feature extraction, which makes it more appropriate and accurate. Thereafter, the discriminative features are re extracted to according to InfoNCE loss based on previous positive and negative samples, which will make the intra-class samples more compact and the inter-class samples more dispersed. At the same time, using the potential structure information of subspace samples to dynamically construct positive and negative samples can make our framework more robust to noisy data. Furthermore, CL-FEFA considers the mutual information between positive samples, that is, similar samples in potential structures, which provides theoretical support for its advantages in feature extraction. The final numerical experiments prove that the proposed framework has a strong advantage over the traditional feature extraction methods and contrastive learning methods.