arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2107.01776cs.CVcs.AI
收藏

基于连续对比学习的图像分类

Continual Contrastive Learning for Image Classification

Peking University(北京大学)

Zhiwei Lin, Yongtao Wang, Hongxiang Lin

摘要:

近年来,自监督表示学习在多媒体技术中得到了进一步的发展。大多数现有的自监督学习方法适用于打包数据。然而,当涉及到流数据时,它们正遭受灾难性遗忘问题的困扰,这一问题没有得到广泛的研究。在本文中,我们首次尝试解决主流自监督方法中的灾难性遗忘问题,即,对比学习方法。具体来说,我们首先开发了一个排练为基础的框架结合了一种新的采样策略和自我监督的知识蒸馏,随着时间的推移有效地传递信息。然后,我们提出了一个额外的样本队列,以帮助网络在嵌入空间中分离新旧数据的特征表示。实验结果表明,在10步增量设置下,与不采用任何技术逐个学习任务的朴素自监督基线相比,在CIFAR-100、ImageNet-Sub和ImageNet-Full上,图像分类准确率分别提高了1.60%、2.86%和1.29%。我们的代码将在https://github.com/VDIGPKU/ContinualContrastiveLearning上提供。

英文摘要:

Recently, self-supervised representation learning gives further development in multimedia technology. Most existing self-supervised learning methods are applicable to packaged data. However, when it comes to streamed data, they are suffering from a catastrophic forgetting problem, which is not studied extensively. In this paper, we make the first attempt to tackle the catastrophic forgetting problem in the mainstream self-supervised methods, i.e., contrastive learning methods. Specifically, we first develop a rehearsal-based framework combined with a novel sampling strategy and a self-supervised knowledge distillation to transfer information over time efficiently. Then, we propose an extra sample queue to help the network separate the feature representations of old and new data in the embedding space. Experimental results show that compared with the naive self-supervised baseline, which learns tasks one by one without taking any technique, we improve the image classification accuracy by 1.60% on CIFAR-100, 2.86% on ImageNet-Sub, and 1.29% on ImageNet-Full under 10 incremental steps setting. Our code will be available at https://github.com/VDIGPKU/ContinualContrastiveLearning.

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。

↑