Side-channel analysis concerns cryptanalytic attacks that exploit the physical environment of cryptographic modules by analyzing leaking emanations and obtaining the encryption key. We propose a multiple-input, multilayer-perceptron-based method for classifying power traces, in which a probability is assigned to each class, indicating the likelihood that each trace corresponds to the label for further recovering the key.

中文翻译：

---

基于多输入、基于多层感知的侧信道攻击痕迹分类

侧信道分析涉及密码分析攻击，该攻击通过分析泄漏的发射并获取加密密钥来利用密码模块的物理环境。我们提出了一种基于多输入、基于多层感知器的功率轨迹分类方法，其中为每个类别分配了一个概率，表明每个轨迹对应于标签的可能性，以进一步恢复密钥。