首页>
外国专利>
ATTACK-LESS ADVERSARIAL TRAINING FOR ROBUST ADVERSARIAL DEFENSE
ATTACK-LESS ADVERSARIAL TRAINING FOR ROBUST ADVERSARIAL DEFENSE
展开▼
机译:对强大的对抗防御的攻击较少的对抗性培训
展开▼
页面导航
摘要
著录项
相似文献
摘要
Disclosed herein is attack-less adversarial training for robust adversarial defense. The attack-less adversarial training for robust adversarial defense includes the steps of: (a) generating individual intervals (ci) by setting the range of color (C) and then discretizing the range of color (C) by a predetermined number (k); (b) generating one batch from an original image (X) and training a learning model with the batch; (c) predicting individual interval indices (ŷialat) from respective pixels (xi) of the original image (X) by using an activation function; (d) generating a new image (Xalat) through mapping and randomization; and (e) training a convolutional neural network with the image (Xalat) generated in step (d) and outputting a predicted label (Ŷ).
展开▼
机译:本文公开了对强大的对抗性防御的攻击较小的对抗性培训。对强大的抗逆性防御的攻击较少的对抗性培训包括以下步骤:(a)通过设置颜色(c)的范围然后离散化颜色范围(c)来生成各个间隔(c i sub>)(c )通过预定数量(k); (b)从原始图像(x)生成一批并使用批次培训学习模型; (c)通过使用使用a的各个像素(x i sub>)预测各个间隔指数(ŷ i sub> alat sup>)激活功能; (d)通过映射和随机化生成新图像(x alat sup>); (e)培训具有在步骤(d)中生成的图像(x alat sup>)并输出预测标签(ŷ)的卷积神经网络。
展开▼