Scope: Deep Learning-Based Side-Channel Evaluation Across Classical And Post-Quantum Cryptographic Implementations

Author

• Nouf Nur Nabilah, Prairie View A&M University

Date of Award

8-2026

Document Type

Thesis

Degree Name

Master of Science

Department

Electrical Engineering

Abstract

A cryptographic algorithm that is mathematically secure can still be broken since its hardware implementation leaks secret information through power consumption or electromagnetic emissions. These vulnerabilities are applicable to widely deployed classical encryption schemes as well as emerging post-quantum schemes which are standardized to provide quantum resistance. This thesis evaluated how deep learning changes such side-channel attacks across both through a unified study of four attacks, spanning a classical block cipher and two protected post-quantum schemes, and tested learned methods against the conditions that separated a laboratory result from a practical attack. Profiled deep-learning attacks recovered keys efficiently under favorable laboratory conditions. However, their behavior was far less well understood when traces were misaligned, when the attacked key differed from the one used in training, and when the leakage was faint and measured on hardware. The thesis asked whether learned attacks remained effective under these conditions, and where they reached their limits.

The four attacks were evaluated on physical devices, targeting AES-128 on a microcontroller and the post-quantum schemes ML-KEM and HQC on field-programmable gate arrays. Each attack was measured by guessing entropy and the number of traces needed to recover a key. On the classical cipher, an ensemble of convolutional networks stabilized recovery against a timing countermeasure. Selecting the most informative trace samples then produced a far smaller model that recovered keys it had never seen, which established that the exploited leakage was a property of the device rather than of any single key. On the post-quantum targets, the limiting factor was not the sophistication of the model but the distance between training and deployment. The smallest network often outperformed an attention-based transformer many times its size, and the gap between simulated and real measurements bounded every learned attack. On a protected implementation whose leak defeated standard attacks, knowledge distillation nonetheless recovered the targeted secret byte. Across the four studies, Deep Learning helped the most when a method was aimed at a specific obstacle, and the least when complexity was added for its own sake proving that the distance between training and deployment limited a practical attack.

Index Terms—Advanced Encryption Standard (AES), deep learning, electromagnetic analysis, Hamming Quasi-Cyclic (HQC), knowledge distillation, Module-Lattice-Based Key-Encapsulation Mechanism (ML-KEM), post-quantum cryptography, power analysis, side-channel analysis.

Committee Chair/Advisor

Annamalai Annamalai

Committee Member

Akshay Raghavendra Kulkarni

Committee Member

Avra Bandyopadhyay

Committee Member

Sheikh Tareq Ahmed

Publisher

Prairie View A&M University

Rights

Date of Digitization

9/14/2026

Contributing Institution

J B Coleman Library

City of Publication

Prairie View

MIME Type

Application/PDF

Recommended Citation

Nabilah, N. (2026). Scope: Deep Learning-Based Side-Channel Evaluation Across Classical And Post-Quantum Cryptographic Implementations. Retrieved from https://digitalcommons.pvamu.edu/pvamu-theses/1686