通过等变表示进行对比学习
Contrastive Learning Via Equivariant Representation
AI总结：
本文提出CLeVER等变对比学习框架，通过兼容任意复杂增强策略提取并融入等变信息，有效提升了主流CL模型的训练效率、鲁棒性及下游任务性能，并增强了旋转不变性与敏感性。
摘要：
不变对比学习（ICL）方法在各个领域取得了令人印象深刻的性能。然而，潜在空间中缺乏关于失真（增强）相关信息的表示，使得ICL在训练效率和下游任务的鲁棒性方面表现次优。最近的研究表明，将等变性引入对比学习（CL）可以提高整体性能。在本文中，我们重新审视了增强策略和等变性在提高CL有效性方面的作用。我们提出了CLeVER（通过等变表示进行对比学习），一种新颖的等变对比学习框架，兼容各种主流CL骨干模型的任意复杂度的增强策略。实验结果表明，CLeVER有效地从实际自然图像中提取并结合等变信息，从而提高了基线模型在下游任务中的训练效率和鲁棒性，并实现了最先进的（SOTA）性能。此外，我们发现利用CLeVER提取的等变信息同时增强了实验任务中的旋转不变性和敏感性，并有助于在处理复杂增强时稳定框架，特别是对于小规模骨干模型。
英文摘要：
Invariant Contrastive Learning (ICL) methods have achieved impressive performance across various domains. However, the absence of latent space representation for distortion (augmentation)-related information in the latent space makes ICL sub-optimal regarding training efficiency and robustness in downstream tasks. Recent studies suggest that introducing equivariance into Contrastive Learning (CL) can improve overall performance. In this paper, we revisit the roles of augmentation strategies and equivariance in improving CL's efficacy. We propose CLeVER (Contrastive Learning Via Equivariant Representation), a novel equivariant contrastive learning framework compatible with augmentation strategies of arbitrary complexity for various mainstream CL backbone models. Experimental results demonstrate that CLeVER effectively extracts and incorporates equivariant information from practical natural images, thereby improving the training efficiency and robustness of baseline models in downstream tasks and achieving state-of-the-art (SOTA) performance. Moreover, we find that leveraging equivariant information extracted by CLeVER simultaneously enhances rotational invariance and sensitivity across experimental tasks, and helps stabilize the framework when handling complex augmentations, particularly for models with small-scale backbones.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
