arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2512.02152cs.CV
收藏

上下文增强对比损失:在对比学习框架中增强样本间内在联系的呈现

Context-Enriched Contrastive Loss: Enhancing Presentation of Inherent Sample Connections in Contrastive Learning Framework

Department of Electrical and Computer Engineering, Western University(电气与计算机工程系,西部大学);Vector Institute(向量研究所)

Haojin Deng, Yimin Yang

AI总结:

本文提出上下文增强对比损失,通过改进学习效果和解决信息失真问题,提升对比学习在大规模数据集中的性能和效率。

摘要:

对比学习在多个大规模基准测试中取得了 popularity,并推动了 state-of-the-art 性能。在对比学习中,对比损失函数在通过旋转或裁剪等技术来辨别样本之间的相似性方面起着关键作用。然而,这种学习机制也可能从增强样本中引入信息失真。这是因为训练后的模型可能会对具有相同标签的样本信息产生显著的过度依赖,同时忽视源自同一初始图像的正样本对,尤其是在大规模数据集的情况下。本文提出了一种上下文增强的对比损失函数,通过包含两个收敛目标同时提高学习效果并解决信息失真问题。第一部分,对标签对比敏感,区分相同类和不同类的特征,从而提高对比训练效率。同时,第二部分将同一源图像的增强样本拉近,并将所有其他样本推远。我们评估了所提出的方法在图像分类任务上的性能,这些任务是大多数广泛接受的8个大规模基准测试数据集:CIFAR10、CIFAR100、Caltech-101、Caltech-256、ImageNet、BiasedMNIST、UTKFace和CelebA数据集。实验结果表明,所提出的方法在泛化性能和学习收敛速度方面优于16种最先进的对比学习方法。有趣的是,我们的技术在解决系统失真任务上表现突出。在下游BiasedMNIST数据集上,相比原始对比损失函数,我们的技术实现了22.9%的改进,凸显了其在更高效和公平的下游训练中的潜力。

英文摘要:

Contrastive learning has gained popularity and pushes state-of-the-art performance across numerous large-scale benchmarks. In contrastive learning, the contrastive loss function plays a pivotal role in discerning similarities between samples through techniques such as rotation or cropping. However, this learning mechanism can also introduce information distortion from the augmented samples. This is because the trained model may develop a significant overreliance on information from samples with identical labels, while concurrently neglecting positive pairs that originate from the same initial image, especially in expansive datasets. This paper proposes a context-enriched contrastive loss function that concurrently improves learning effectiveness and addresses the information distortion by encompassing two convergence targets. The first component, which is notably sensitive to label contrast, differentiates between features of identical and distinct classes which boosts the contrastive training efficiency. Meanwhile, the second component draws closer the augmented samples from the same source image and distances all other samples. We evaluate the proposed approach on image classification tasks, which are among the most widely accepted 8 recognition large-scale benchmark datasets: CIFAR10, CIFAR100, Caltech-101, Caltech-256, ImageNet, BiasedMNIST, UTKFace, and CelebA datasets. The experimental results demonstrate that the proposed method achieves improvements over 16 state-of-the-art contrastive learning methods in terms of both generalization performance and learning convergence speed. Interestingly, our technique stands out in addressing systematic distortion tasks. It demonstrates a 22.9% improvement compared to original contrastive loss functions in the downstream BiasedMNIST dataset, highlighting its promise for more efficient and equitable downstream training.

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。

↑