Dropping Pixels for Adversarial Robustness

机译：抗逆性鲁棒性的像素

获取原文

页面导航

摘要
著录项
相似文献
相关主题

摘要

Deep neural networks are vulnerable against adversarial examples. In this paper, we propose to train and test the networks with randomly subsampled images with high drop rates. We show that this approach significantly improves robustness against adversarial examples in all cases of bounded L0, L2 and L∞ perturbations, while reducing the standard accuracy by a small value. We argue that subsampling pixels can be thought to provide a set of robust features for the input image and, thus, improves robustness without performing adversarial training.

机译：深度神经网络易受对抗性的例子。在本文中，我们建议使用具有高滴加率的随机限速图像来培训和测试网络。我们表明，在所有有界L0，L2和L∞扰动的情况下，这种方法显着改善对抗对抗的实例的鲁棒性，同时通过小值降低标准精度。我们认为可以认为可以认为输入图像的一组鲁棒特征，因此，在不进行对抗性训练的情况下提高鲁棒性。

著录项

来源
《IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops》|2019年|1 v.|共7页
会议地点
作者
Hossein Hosseini; Sreeram Kannan; Radha Poovendran;
展开▼
作者单位

展开▼
会议组织
原文格式 PDF
正文语种
中图分类 TP391.41;
关键词
Robustness; Training; Standards; Image recognition; Perturbation methods; Correlation; Predictive models;

机译：鲁棒性;培训;标准;图像识别;扰动方法;相关性;预测模型;

相似文献

外文文献
中文文献
专利

1. Improving the Quality of Synthetic FLAIR Images with Deep Learning Using a Conditional Generative Adversarial Network for Pixel-by-Pixel Image Translation [J] . Hagiwara A., Otsuka Y., Hori M., AJNR. American journal of neuroradiology . 2019,第2期

机译：利用条件生成对抗网络提高与深层学习的合成风格图像的质量，用于逐像素图像转换
2. Robust Graph Neural Networks Against Adversarial Attacks via Jointly Adversarial Training [J] . Hu Tian, Bowei Ye, Xiaolong Zheng, IFAC PapersOnLine . 2020,第5期

机译：通过共同对抗培训，强大的图形神经网络对抗对抗攻击
3. Reinforcing Adversarial Robustness using Model Confidence Induced by Adversarial Training [J] . Xi Wu, Uyeong Jang, Jiefeng Chen, JMLR: Workshop and Conference Proceedings . 2018,第3期

机译：利用对抗训练产生的模型置信度增强对抗鲁棒性
4. Dropping Pixels for Adversarial Robustness [C] . Hossein Hosseini, Sreeram Kannan, Radha Poovendran IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops . 2019

机译：降低像素的对抗性
5. Towards Adversarial and Non-Adversarial Robustness of Machine Learning and Signal Processing: Fundamental Limits and Algorithms [D] . Yi, Jirong. 2021

机译：对机器学习和信号处理的侵犯和非对抗性鲁棒性：基本限制和算法
6. Improving the Quality of Synthetic FLAIR Images with Deep Learning Using a Conditional Generative Adversarial Network for Pixel-by-Pixel Image Translation [O] . A. Hagiwara, Y. Otsuka, M. Hori, 2019

机译：使用条件生成对抗网络进行逐像素图像翻译的深度学习提高合成FLAIR图像的质量
7. Dropping Pixels for Adversarial Robustness [O] . Hossein Hosseini, Sreeram Kannan, Radha Poovendran 2019

机译：抗逆性鲁棒性的像素

Dropping Pixels for Adversarial Robustness

摘要

著录项

相似文献

相关主题

期刊订阅