Contrastive learning has emerged as an essential approach for self-supervised learning in computer vision. The central objective of contrastive learning is to maximize the similarities between two augmented versions of the same image (positive pairs), while minimizing the similarities between different images (negative pairs). Recent studies have demonstrated that harder negative samples, i.e., those that are difficult to distinguish from anchor sample, play a more critical role in contrastive learning. In this paper, we propose a novel featurelevel method, namely sampling synthetic hard negative samples for contrastive learning (SSCL), to exploit harder negative samples more effectively. Specifically, 1) we generate more and harder negative samples by mixing negative samples, and then sample them by controlling the contrast of anchor sample with the other negative samples. 2) Considering that the negative samples obtained by sampling may have the problem of false negative samples, we further debias the negative samples. Our proposed method improves the classification performance on different image datasets and can be readily applied to existing methods.

翻译：对比学习已成为计算机视觉自监督学习中的关键方法。其核心目标是在最大化同一图像两个增强版本（正样本对）之间相似性的同时，最小化不同图像（负样本对）之间的相似性。近期研究表明，更难区分的负样本（即与锚点样本难以辨别的样本）在对比学习中起更关键的作用。本文提出一种新颖的特征层面方法——采样合成难负样本用于对比学习（SSCL），以更有效地利用难负样本。具体而言：1）通过混合负样本生成更多且更难的负样本，并控制锚点样本与其他负样本之间的对比度进行采样；2）针对采样所得负样本可能存在的假负样本问题，进一步对负样本进行去偏处理。所提方法在不同图像数据集上提升了分类性能，并可便捷地应用于现有方法。