本文是关于侧信道攻击入门的指北，不适合未满8岁的人士食用，未成年人请在成年人的陪同下浏览。本人对指北内容概不负责，如有眩晕、呕吐等不适情况请立刻关闭该网页并及时就医。

论文（持续补充）

这里是本人看过（下载过）的论文，会持续补充。

# 题目 站内链接 论文链接 类型
0 Study of Deep Learning Techniques for Side-Channel Analysis and Introduction to ASCAD Database click link 入门必看
1 One Network to rule them all. An autoencoder approach to encode datasets click link 特征提取
2 Multilabel Deep Learning-Based Side-Channel Attack click link 多标签学习
3 Deep Learning based Side-Channel Attack: a New Profiling Methodology based on Multi-Label Classification click link 多标签学习
4 CDAE: Towards Empowering Denoising in Side-Channel Analysis click link 降噪
5 On the Performance of Convolutional Neural Networks for Side-channel Analysis click link 网络架构
6 To Overfit, Or Not to Overfit: Improving the Performance of Deep Learning-based SCA click link 炼丹指导
7 The Best of Two Worlds: Deep Learning-assisted Template Attack click link 特征提取方法
8 Pay Attention to Raw Traces: A Deep Learning Architecture for End-to-End Profiling Attacks link 端到端侧信道攻击
9 Cross-Device Profiled Side-Channel Attack with Unsupervised Domain Adaptation link 跨设备侧信道攻击
10 Methodology for Efficient CNN Architectures in Profiling Attacks link 网络架构
11 It’s a Kind of Magic: A Novel Conditional GAN Framework for Efficient Profiling Side-channel Analysis link 特征提取
12 I Know What Your Layers Did: Layer-wise Explainability of Deep Learning Side-channel Analysis link
13 Revisiting a Methodology for Efficient CNN Architectures in Profiling Attacks click link 练手用
14 Don’t Learn What You Already Know Scheme-Aware Modeling for Profiling Side-Channel Analysis against Masking link 没看懂
15 Exploring Feature Selection Scenarios for Deep Learning-based Side-channel Analysis click link 特征工程
16 Side-channel analysis against ANSSI’s protected AES implementation on ARM: end-to-end attacks with multi-task learning link 多任务学习
17 Multi-Leak Deep-Learning Side-Channel Analysis link 多点泄露攻击
18 Deep Learning Side-Channel Collision Attack link 碰撞攻击（新
19 Remove Some Noise: On Pre-processing of Side-channel Measurements with Autoencoders click link 降噪
20 THE CIRCLE OF DL-SCA — link 吴立超博士毕业论文
21 I Choose You: Automated Hyperparameter Tuning for Deep Learning-based Side-channel Analysis link 超参数自动调整
22 Hiding in Plain Sight: Non-profiling Deep Learning-based Side-channel Analysis with Plaintext/Ciphertext link 非建模侧信道攻击
23 Breaking Cryptographic Implementations Using Deep Learning Techniques
24 SoK: Deep Learning-based Physical Side-channel Analysis link 综述
25 Fake it till you make it: Data Augmentation using Generative Adversarial Networks for all the crypto you need on small devices link 数据增强
26
27 The Need for Speed: A Fast Guessing Entropy Calculation for Deep Learning-Based SCA link 训练中callback
28 Ranking Loss: Maximizing the Success Rate in Deep Learning Side-Channel Analysis link 损失函数
29 Label Correlation in Deep Learning-Based Side-Channel Analysis link 标签相关性
30 Breaking Free: Leakage Model-free Deep Learning-based Side-channel Analysis link 泄露模型
31 Pay Attention to Raw Traces: A Deep LearningArchitecture for End-to-End Profiling Attacks link 攻击原始能量迹
32 Deep Learning Side-Channel Analysis on Large-Scale Traces A Case Study on a Polymorphic AES link 攻击大规模能量迹
33 Non-Profiled Deep Learning-based Side-Channel attacks with Sensitivity Analysis link 敏感值分析
34 EstraNet: An Efficient Shift-Invariant Transformer Network for Side-Channel Analysis link Trasformer 侧信道
35 A Comprehensive Study of Deep Learning for Side-Channel Analysis — link NLL
36 Leakage Certification Revisited: Bounding Model Errors in Side-Channel Security Evaluations link 感知信息
37 Perceived Information Revisited New Metrics to Evaluate Success Rate of Side-Channel Attacks link 感知信息pro
38 Perceived Information Revisited II Information-Theoretical Analysis of Deep-Learning Based Side-Channel Attacks link 感知信息pro max
39 AutoPOI: automated points of interest selection for side-channel analysis link 自动提取兴趣点
40 Weakly Profiling Side-channel Analysis link 新建模方法
41 Plaintext: A Missing Feature for Enhancing the Power of Deep Learning in Side-Channel Analysis? Breaking multiple layers of side-channel countermeasures link 提升针对掩码的破坏性
42 Strength in Numbers: Improving Generalization with Ensembles in Machine Learning-based Profiled Side-channel Analysis link 集成学习
43 Give Me 5 Minutes Attacking ASCAD with a Single Side-Channel Trace link SCALib的广告
44 Mind the Portability: A Warriors Guide through Realistic Profiled Side-channel Analysis link 跨设备侧信道攻击
45 Ablation Analysis for Multi-Device Deep Learning-Based Physical Side-Channel Analysis link 跨设备侧信道攻击
46 Adaptive Chosen-Plaintext Deep Learning-based Side-Channel Analysis link
47 How Far Can We Reach? Breaking RSM-Masked AES-128 Implementation link RSM掩码攻击
48 Sneaking up the Ranks: Partial Key Exposure Attacks on Rank-Based Schemes link rank
49 Manifold Learning Side-Channel Attacks against Masked Cryptographic Implementations link 流形学习侧信道
50 Single-trace side-channel attacks on MAYO exploiting leaky modular multiplication link 后量子侧信道攻击
51 One Solves All: Exploring ChatGPT’s Capabilities for Fully Automated Simple Power Analysis on Cryptosystems link GPT-SPA
52 Can KANs Do It? Toward Interpretable Deep Learning-based Side-channel Analysis link KAN-SCA
53 Creating from Noise: Trace Generations Using Diffusion Model for Side-Channel Attack link 扩散模型-sca
54 Leakage Model-flexible Deep Learning-based Side-channel Analysis link 多任务多标签侧信道攻击
55 Scoring the predictions: a way to improve profiling side-channel attacks link 排名优化
56 Regularizers to the rescue: fighting overfitting in deep learning-based side-channel analysis link 过拟合优化
58 CL-SCA: Leveraging Contrastive Learning for Profiled Side-Channel Analysis link 对比学习
60 Autoencoder-enabled model portability for reducing hyperparameter tuning efforts in side-channel analysis link 超参数调整优化策略
61 Diffuse Some Noise: Diffusion Models for Measurement Noise Removal in Side-channel Analysis link 扩散模型去噪
62 Blind-Folded: Simple Power Analysis Attacks using Data with a Single Trace and no Training link 单条能量迹、SPA
63 Side-Channel Attacks Based on Multi-Loss Regularized Denoising AutoEncoder [link](Side-Channel Attacks Based on Multi-Loss Regularized Denoising AutoEncoder) 去噪编码器
64 HierNet: A Hierarchical Deep Learning Model for SCA on Long Traces link 原始能量迹攻击
65 On the Instability of Softmax Attention-Based Deep Learning Models in Side-Channel Analysis link 注意力
66 Conditional Variational AutoEncoder based on Stochastic Attacks link 条件变分自编码器
67 Beyond the Last Layer: Deep Feature Loss Functions in Side-channel Analysis link 损失函数
68 OccPoIs: Points of Interest based on Neural Network’s Key Recovery in Side-Channel Analysis through Occlusion link
69 Peek into the Black-Box: Interpretable Neural Network using SAT Equations in Side-Channel Analysis link 可解释性研究
70 An End-to-end Plaintext-based Side-channel Collision Attack without Trace Segmentation link 碰撞攻击
71 SCA-CGAN:A New Side-Channel Attack Method for Imbalanced Small Samples link 小样本侧信道攻击
72 The Need for MORE: Unsupervised Side-channel Analysis with Single Network Training and Multi-output Regression link 无监督侧信道攻击
73 Tipping the Balance: Imbalanced Classes in Deep Learning Side-channel Analysis link 标签不平衡研究
74 The Side-channel Metrics Cheat Sheet link 好文！
75 No (good) loss no gain: systematic evaluation of loss functions in deep learning-based side-channel analysis link 损失函数设计
76 It’s a Kind of Magic: A Novel Conditional GAN Framework for Efficient Profiling Side-channel Analysis link GAN
77 Efficient Exploitation of Noise Leakage for Template Attack link 模板攻击
78 Encoding Power Traces as Images for Efficient Side-Channel Analysis link 图像侧信道
79

Mathematical Foundations for Side-Channel Analysis of Cryptographic Systems 侧信道攻击的数学基础，但是学校没springer的Institutional subscriptions！！！

指北

这部分解释侧信道攻击以及本人研究方向的内容。

什么是侧信道攻击？

在我小的时候，经常趁父母出去上班不在家的空挡偷偷打开电视，看一些古侠剧（天龙八部、神雕侠侣…），里边的各种情节都十分吸引我。但是，我无时无刻不得不关注的就是门外的脚步声，那个时候只要二老回家，脚步声听的是清清楚楚，直接电视关机飞到小桌子前假装写作业一气呵成。没想到小小年纪便用到了我高贵的侧信道攻击（Side Channel Attack，SCA）技术！也就是说我并不需要看到他俩这个主体，通过上楼梯时所泄露出来的脚步声这一侧信息就可以判别是不是他们。

更令我没想到的是，后来他们也用到了侧信道分析！并不需要看到我在看电视，而是仅仅需要摸一下电视的大脑袋，通过电视机壳传出来的温度侧信息，去决定是否要对我发起物理攻击（不是。

在密码学领域中，侧信道攻击并不是直接针对的密码的算法层面，相较于其他理论上的分析攻击方法，该方法直接对设备加密时所泄漏的侧信息进行分析。例如在RSA密码算法中，当加密算法执行到求平方乘步骤时，指数位为0与为1所需要的执行耗时是不同的，攻击者可以通过观察执行泄露的时间侧信息来进行处理，进而恢复密钥。

侧信道攻击又分为建模类侧信道攻击(Profiling Side Channel Attack)以及非建模类侧信道攻击(Non-Profiling Side Channel Attack)。建模类侧信道攻击要求攻击者事先拥有一个与被攻击设备相同或相似的副本，通过采集在副本设备上执行密码算法所泄露出来的侧信息并加以分析建模。最后，将建好的模型用在被攻击设备上所泄漏的侧信息中，以完成关键信息的恢复，常见的攻击方法有模板攻击、随机模型攻击以及基于深度学习的侧信道攻击。非建模类侧信道攻击则直接对被攻击设备发起攻击，常见的攻击方法有SPA、DPA、CPA。

如何入门侧信道攻击

由于本人研究方向为侧信道攻击与深度学习的交叉学科（Deep Learning based Side Channel Attack，DLSCA），所以在学习相关知识的时候更偏向于深度学习建模的方法，大致分为兴趣点选择、数据预处理、特征工程、模型的训练、评估五部分[1]，具体如图1所示。

图1 基于深度学习的侧信道攻击流程

以我个人经验，需要先搞懂的是你要建模的侧信息是什么，以下是在攻击时常用的数据集。每个数据集中都包含了大量的能量迹，每条能量迹又存在对应的时间样本点。之后的事情就是看论文做实验，下一部分有一些可以直接上手的基础实验，SCA_Box中的实验大家在做的时候如果有疑问可以直接在上边提Issue或Email联系我，我看到了都会回复。

数据集 🔗链接
ASCAD https://github.com/ANSSI-FR/ASCAD/tree/master
CHESCTF_2018 http://aisylabdatasets.ewi.tudelft.nl/ches_ctf.h5
AES_HD http://aisylabdatasets.ewi.tudelft.nl/aes_hd_ext.npz
AES_RD https://github.com/ikizhvatov/randomdelays-traces
DPAv4.2_Zaid https://github.com/gabzai/Methodology-for-efficient-CNN-architectures-in-SCA/blob/master/DPA-contest v4/DPAv4_dataset.zip

需要复现的实验（有源码）

目前，DLSCA的论文有很多都给出了代码，大家有时间都要复现一下。我在下边贴出来论文以及对应代码链接，还有一些基础实验的复现代码，可以结合自己情况审代码以及实现。

基础实验：

论文及代码

期刊

期刊 链接🔗 评级
TCHES https://tches.iacr.org/ 顶刊⭐️⭐️⭐️⭐️+
JCEN 此生必投！！！ https://link.springer.com/journal/13389 中科院四区 JCR Q2 ⭐️⭐️⭐️
ePrint https://eprint.iacr.org/ 预印本（占坑）
TIFS https://ieeexplore.ieee.org/xpl/RecentIssue.jsp?punumber=10206 中科院一区 JCRQ1 ⭐️⭐️⭐️⭐️⭐️
Joc ⭐️⭐️⭐️⭐️⭐️

会议

会议 链接
ICCD https://www.iccd-conf.com/Home.html

参考文献

  1. SoK: Deep Learning-based Physical Side-channel Analysis ↩︎