Revisiting Contrastive Learning through the Lens of Neighborhood Component Analysis: an Integrated Framework
Revisiting Contrastive Learning through the Lens of Neighborhood Component Analysis: an Integrated Framework
复制标题
DOI：
--
复制
发表时间：
2021-12
期刊：
影响因子：
--
通讯作者：
Ching-Yun Ko;Jeet Mohapatra;Sijia Liu;Pin-Yu Chen;Lucani E. Daniel;Lily Weng
中科院分区：
文献类型：
--
作者：
Ching-Yun Ko;Jeet Mohapatra;Sijia Liu;Pin-Yu Chen;Lucani E. Daniel;Lily Weng
对比学习作为自监督表征学习的重要工具，近年来得到了前所未有的关注。本质上，对比学习旨在利用正负样本对进行表示学习，这涉及到利用特征空间中的邻域信息。通过研究对比学习和邻域成分分析（NCA）之间的联系，提出了一种新的对比学习的随机最近邻观点，并提出了一系列优于现有的对比损失.在我们提出的框架下，我们展示了一种新的方法来设计综合对比损失，可以同时实现良好的准确性和鲁棒性的下游任务。通过集成框架，我们的标准准确度提高了6%，鲁棒准确度提高了17%。
As a seminal tool in self-supervised representation learning, contrastive learning has gained unprecedented attention in recent years. In essence, contrastive learning aims to leverage pairs of positive and negative samples for representation learning, which relates to exploiting neighborhood information in a feature space. By investigating the connection between contrastive learning and neighborhood component analysis (NCA), we provide a novel stochastic nearest neighbor viewpoint of contrastive learning and subsequently propose a series of contrastive losses that outperform the existing ones. Under our proposed framework, we show a new methodology to design integrated contrastive losses that could simultaneously achieve good accuracy and robustness on downstream tasks. With the integrated framework, we achieve up to 6\% improvement on the standard accuracy and 17\% improvement on the robust accuracy.