發(fā)簡(jiǎn)信
IP屬地：廣東
-
Adversarial Distributional Training
Adversarial Distributional Training for Robust Deep LearningZhijie Deng,...
-
定制化對(duì)抗訓(xùn)練Customized Adversarial TrainingCAT: Customized Adversarial Training for Improved RobustnessMinhao Cheng...
-
聚類(lèi)訓(xùn)練Clustering Training for Robustness
ClustTR: Clustering Training for RobustnessMotasem Alfarra, Juan C. Pére...
-
對(duì)抗樣本領(lǐng)域綜述文章推薦
對(duì)于剛接觸對(duì)抗樣本領(lǐng)域的小伙伴們來(lái)說(shuō)，看到領(lǐng)域內(nèi)眾多文章時(shí)，簡(jiǎn)直眼花繚亂。這時(shí)候，如果一篇好的綜述概括了當(dāng)前領(lǐng)域內(nèi)的主要進(jìn)展，提供給我們?cè)擃I(lǐng)域的...
-
[論文筆記](méi)DeepFool: a simple and accurate method to fool deep neural networks題目：DeepFool: a simple and accurate method to fool deep neural networks地址...
-
[論文筆記](méi)Towards Evaluating the Robustness of Neural Networks
題目：Towards Evaluating the Robustness of Neural Networks地址：https://arxiv....
-
[論文閱讀筆記](méi)One pixel attack for fooling deep neural networks
論文題目：One pixel attack for fooling deep neural networks論文地址：https://arxiv...
-
對(duì)抗攻擊方法匯總（持續(xù)更新）
自從2014年Szegedy等人提出對(duì)抗樣本以來(lái)，不斷有研究者提出新的對(duì)抗攻擊方法。本文匯總了當(dāng)前已有的絕大多數(shù)算法，以拋磚引玉用，并不斷更新。...
-
[論文閱讀筆記](méi)The Limitations of Deep Learning in Adversarial Settings論文題目：The Limitations of Deep Learning in Adversarial Settings論文地址：https:...