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 2008.02427cs.CVcs.MM
收藏

从噪声数据中回收可重用样本用于鲁棒学习

Salvage Reusable Samples from Noisy Data for Robust Learning

Nanjing University of Science and Technology(南京理工大学);Alibaba Group(阿里巴巴集团);AnyVision;University of Technology Sydney(悉尼科技大学)

Zeren Sun, Xian-Sheng Hua, Yazhou Yao, Xiu-Shen Wei, Guosheng Hu, Jian Zhang

摘要:

由于Web图像中标签噪声的存在和深度神经网络的高记忆能力,直接通过Web图像训练深度细粒度(FG)模型往往具有较差的识别能力。在文献中,为了缓解这个问题,损失校正方法试图估计噪声转移矩阵,但不可避免的错误校正会导致严重的累积误差。样本选择方法基于小的损失可以减轻累积误差的事实来识别干净(“容易”)样本。然而,可以提高FG模型鲁棒性的“硬”和错误标记的示例也被丢弃。为此,我们提出了一种基于确定性的可重用样本选择和校正方法,称为CRSSC,用于处理使用Web图像训练深度FG模型时的标签噪声。我们的主要想法是额外识别和纠正可重用的样本,然后利用它们与干净的示例一起更新网络。我们从理论和实验的角度证明了所提出的方法的优越性。

英文摘要:

Due to the existence of label noise in web images and the high memorization capacity of deep neural networks, training deep fine-grained (FG) models directly through web images tends to have an inferior recognition ability. In the literature, to alleviate this issue, loss correction methods try to estimate the noise transition matrix, but the inevitable false correction would cause severe accumulated errors. Sample selection methods identify clean ("easy") samples based on the fact that small losses can alleviate the accumulated errors. However, "hard" and mislabeled examples that can both boost the robustness of FG models are also dropped. To this end, we propose a certainty-based reusable sample selection and correction approach, termed as CRSSC, for coping with label noise in training deep FG models with web images. Our key idea is to additionally identify and correct reusable samples, and then leverage them together with clean examples to update the networks. We demonstrate the superiority of the proposed approach from both theoretical and experimental perspectives.

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。

↑