This document describes AI security and privacy protection. As a universal technology, AI brings huge opportunities and benefits, but also faces new security and privacy protection challenges. MindArmour is a sub-project of MindSpore. It provides security and privacy protection capabilities for MindSpore, including technologies such as adversarial robustness, model security test, differential privacy training, and privacy breach risk assessment.
The Attack
base class defines the interface for generating adversarial examples. Its subclasses implement various specific generation algorithms and allow security personnel to quickly and efficiently generate adversarial examples for attacking AI models to evaluate the robustness of the models.
The Defense
base class defines the interface for adversarial training. Its subclasses implement various adversarial training algorithms to enhance the adversarial robustness of the models.
The Detector
base class defines the interface for adversarial sample detection. Its subclasses implement various specific detection algorithms to enhance the adversarial robustness of the models.
For details, see Improving Model Security with NAD Algorithm.
The Fuzzer
class controls the fuzzing process based on the neuron coverage gain. It uses natural perturbation and adversarial sample generation methods as the mutation policy to activate more neurons to explore different types of model output results and error behavior, helping users enhance model robustness.
For details, see Testing Model Security Using Fuzz Testing.
DPModel
inherits mindspore.Model
and provides the entry function for differential privacy training.
For details, see Protecting User Privacy with Differential Privacy Mechanism.
SuppressModel
inherits mindspore.Model
and provides the entry function for suppress privacy training.
For details, see Protecting User Privacy with Suppress Privacy Mechanism.
The MembershipInference
class provides a reverse analysis method. It can infer whether a sample is in the training set of a model based on the prediction information of the model on the sample to evaluate the privacy breach risk of the model.
For details, see Testing Model Security with Membership Inference.
此处可能存在不合适展示的内容,页面不予展示。您可通过相关编辑功能自查并修改。
如您确认内容无涉及 不当用语 / 纯广告导流 / 暴力 / 低俗色情 / 侵权 / 盗版 / 虚假 / 无价值内容或违法国家有关法律法规的内容,可点击提交进行申诉,我们将尽快为您处理。