AUT Journal of Electrical Engineering

AUT Journal of Electrical Engineering

Breaking Probabilistic Side-Channel Defenses: A Deep Learning Approach to Cryptographic Key Recovery

Document Type : Research Article

Authors
Mehrshad Eskandarpour
Parham Soltani
MohammadJavad Jannati
Deparment of Electrical Engineering, IUST, Tehran, Iran.
10.22060/eej.2026.24478.5712
Abstract
Probabilistic dummy operations inject randomized activity into power traces to blur key-dependent leakage, blunting classical side-channel attacks such as CPA and DPA. We introduce a profiling attack that treats traces as sequences of windows and learns to separate key-dependent computation from dummy activity. A lightweight recurrent sequence classifier is trained on traces from an identical device with dummies disabled, producing a model that scores windows for key-bearing work.
At attack time, the classifier filters dummy-protected traces and the retained windows feed a standard likelihood or correlation-based-key-ranking stage. The key-recovery advantage arises because filtering removes windows with negligible key-dependent leakage, increasing the effective signal-to-noise ratio for classical distinguishers, while the recurrent architecture’s temporal context enables robust detection despite timing jitter and variable dummy density. On a DES implementation with randomized dummy insertion, our method attains rank-0 with substantially fewer traces than CPA, DPA and a tuned CNN, and remains robust under timing jitter (±10 samples), varying dummy rates (p = 0.3–0.7), and low SNR(≤5dB). We report window-level metrics (AUC, ) and key-level success curves (rank vs. traces), with ablations isolating the effects of alignment error and dummy probability. The results demonstrate that probabilistic dummy insertion alone is insufficient against sequence-aware profiling attacks, and that hybrid DL-classical pipelines can outperform both pure classical and pure end-to-end deep learning approaches.
Keywords
Side-Channel Analysis
Deep Learning
Power Trace Classification
Dummy Operation Detection
Hardware Security
Countermeasure
Subjects
[1]     M. Ashok, E. V. Levine, and A. P. Chandrakasan, “Randomized switching SAR (RS-SAR) ADC protections for power and electromagnetic side-channel security,” in Proc. IEEE Custom Integrated Circuits Conf. (CICC), 2022.
[2]     A. T. Mozipo and J. M. Acken, “Analysis of countermeasures against remote and local power side-channel attacks using correlation power analysis,” IEEE Trans. Dependable and Secure Computing (Nov. 2024).
[3]     V. V. Gadde, H. Awano, and M. Ikeda, “An encryption-authentication unified A/D conversion scheme for IoT sensor nodes,” in Proc. IEEE Asian Solid-State Circuits Conf. (A-SSCC), 2018.
[4]     T. Miki, N. Miura, H. Sonoda, K. Mizuta, and M. Nagata, “A random interrupt dithering SAR technique for secure ADC against reference-charge side-channel attack,” IEEE Trans. Circuits and Systems II: Express Briefs (Jan. 2020).
[5]     L. Fang, J. Liu, Y. Zhu, C.-H. Chan, and R. P. Martins, “LSB-reused protection technique in secure SAR ADC against power side-channel attack,” in Proc. Asian Hardware Oriented Security and Trust Symp. (AsianHOST), 2022.
[6]     J. Xu and H. M. Heys, “Template attacks of a masked S-box circuit: A comparison between static and dynamic power analyses,” in Proc. IEEE Int. NEWCAS Conf., 2018.
[7]     N. Kaushik and J. Hu, “A switched-capacitor power side-channel attack detection circuit in 65-nm CMOS,” in Proc. IEEE Int. Symp. Circuits and Systems (ISCAS), 2021.
[8]     H. Jeon, N. Karimian, and T. Lehman, “A new foe in GPUs: Power side-channel attacks on neural network,” in Proc. Int. Symp. Quality Electronic Design (ISQED), 2021.
[9]     S. Kumar, S. Chatterjee, C. K. Dabhi, H. Amrouch, and Y. S. Chauhan, “A novel approach to mitigate power side-channel attacks for emerging negative capacitance transistor technology,” in Proc. IEEE Int. NEWCAS Conf., 2022.
[10]  N. Kaushik and J. Hu, “Performance and noise trade-off for SC-based power side-channel attack detection circuits,” in Proc. IEEE Midwest Symp. Circuits and Systems (MWSCAS), 2021.
[11]  G. Perin, L. Wu, and S. Picek, “Exploring feature selection scenarios for deep learning-based side-channel analysis,” IACR Trans. Cryptographic Hardware and Embedded Systems (TCHES), 2022(4).
[12]  S. Picek, G. Perin, L. Mariot, L. Wu, and L. Batina, “SoK: Deep learning-based physical side-channel analysis,” ACM Computing Surveys, 55(11), 2023.
[13]  L. Wu, S. Picek, and L. Batina, “On the evaluation of deep learning-based side-channel analysis,” in Proc. COSADE (LNCS 13211), Springer, 2022.
[14]  G. Zaid, L. Bossuet, F. Dassance, A. Habrard, and A. Venelli, “Ranking Loss: Maximizing the success rate in deep learning side-channel analysis,” IACR Trans. Cryptographic Hardware and Embedded Systems (TCHES), 2021(3).
[15]  M. Kerkhof, M. Czyz, G. Perin, and S. Picek, “No (Good) Loss, No Gain: Systematic evaluation of loss functions in DL-SCA,” Journal of Cryptographic Engineering, 2023.
[16]  L. Wu, G. Perin, and S. Picek, “I Choose You: Automated hyperparameter tuning for deep learning-based side-channel analysis,” IEEE Trans. Emerging Topics in Computing, 12(2), 2024.
[17]  M. Ninan, E. Nimmo, S. Reilly, C. Smith, W. Sun, B. Wang, and J. M. Emmert, “A second look at the portability of deep learning side-channel attacks over EM traces,” in Proc. RAID, 2024.
[18]  C. Wang, M. Ninan, S. Reilly, J. Ward, W. Hawkins, B. Wang, and J. M. Emmert, “Portability of deep-learning side-channel attacks against software discrepancies,” in Proc. ACM WiSec, 2023.
[19]  M. Krček, L. Wu, G. Perin, and S. Picek, “Shift-invariance robustness of CNNs in side-channel analysis,” Mathematics, 12(20), 2024.
[20]  S. Hajra, M. Alam, S. Saha, S. Picek, and D. Mukhopadhyay, “On the instability of softmax attention-based deep learning models in side-channel analysis,” IEEE Trans. Information Forensics and Security, 19, 2024.
[21]  A. Sajadi, N. Zidaric, T. Stefanov, and N. Mentens, “A systematic comparison of side-channel countermeasures for RISC-V-based SoCs,” in Proc. IEEE Nordic Circuits and Systems Conf. (NorCAS), 2024.
[22]  J. Moon, S. Lee, H. Cho, and Y. Oh, “Side-channel analysis for ARIA based on dummy instruction insertion of RISC-V,” Journal of Digital Contents Society (Aug. 2022).
[23]  K. Nomikos, I. K. Siglidis, B. Giese, D. Nagy, N. Kallimopoulos, D. S. Kalogerias, and D. Soudris, “Evaluation of hiding-based countermeasures against DL-SCA with pre-trained networks,” in Proc. IEEE Int. Symp. Defect and Fault Tolerance (DFT), 2022.
[24]  C. Wang, J. Dani, S. Reilly, A. Brownfield, B. Wang, and J. M. Emmert, “TripletPower: Deep-learning side-channel attacks over few traces,” in Proc. IEEE Int. Symp. Hardware Oriented Security and Trust (HOST), 2023.
[25]  K. Pu, J. Yang, X. Tang, J. Feng, and Z. Chai, “A quantitative analysis of non-profiled side-channel attacks with attention,” Electronics, 12(15), 2023.
[26]   L. Wu, G. Perin, and S. Picek, “Weakly profiling side-channel analysis,” IACR Trans. Cryptographic Hardware and Embedded Systems (TCHES), 2024(4).
[27]  L. Wu, G. Perin, and S. Picek, “The best of two worlds: Deep learning-assisted template attack,” IACR Trans. Cryptographic Hardware and Embedded Systems (TCHES), 2022(3).
[28]  Y. Zhu, S. Song, B. Kong, S. Qu, Z. Leng, X. Huang, and L. Liu, “LDL-SCA: Linearized deep learning side-channel attack targeting multi-tenant FPGAs,” in Proc. ACM Great Lakes Symp. VLSI (GLSVLSI), 2024.
[29]  G. Chiari, L. Weissbart, P. Nebel, M. Ammar, M. Magno, and L. Benini, “A deep-learning technique to locate cryptographic operations in side-channel traces,” in Proc. DATE, 2024.
[30]  A. Rezaeezade, T. Yap, D. Jap, S. Bhasin, and S. Picek, “Breaking the blindfold: Deep learning-based blind side-channel analysis,” in Proc. USENIX Security Symp., 2025.
[31]  P. Soltani, M. Eskandarpour, A. Ahmadizad, and H. Soleimani, “Energy-Efficient Routing Algorithm for Wireless Sensor Networks: A Multi-Agent Reinforcement Learning Approach,” arXiv preprint arXiv:2508.14679, 2025. [Online]. Available: https://arxiv.org/abs/2508.14679.
[32]  P. Soltani, M. Eskandarpour, S. Heidari, F. Alizadeh, and H. Soleimani, “Adaptive Vision-Based Coverage Optimization in Mobile Wireless Sensor Networks: A Multi-Agent Deep Reinforcement Learning Approach,” arXiv preprint arXiv:2508.14676, 2025. [Online]. Available: https://arxiv.org/abs/2508.14676.
AUT Journal of Electrical Engineering
Volume 58, Special Issue 1
2026
Pages 113-138