EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Deep-learning side-channel analysis of embedded ML-KEM: Single-trace message recovery, cross-device generalization, and frequency-band leakage characterization

Chung-Wei Kuo iD, Yu-I Hung, Kai-Lun Lin

PLOS One · 04.11.2026 · 21 (11) · e0358974

DOI: 10.1371/journal.pone.0358974 ↗

Kurz gefasst

Fachartikel in der Zeitschrift PLOS One, online erschienen am 04.11.2026. Verfasst von Chung-Wei Kuo und 2 weiteren Autor:innen. Beteiligte Einrichtung: Department of Information Engineering and Computer Science, Feng Chia University, Taichung, Taiwan. Hauptthemen laut MEDLINE: Deep Learning, Computer Security. Die Arbeit nennt 34 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Science and Technology Council.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Quantum computing undermines traditional public-key cryptosystems like RSA and ECC, prompting NIST to standardize Module Lattice Key Encapsulation Mechanism (ML-KEM) as a core Post-Quantum Cryptography (PQC) scheme in August 2024. Despite its theoretical security, ML-KEM’s practical vulnerability in embedded systems remains underexplored. This study evaluates ML-KEM on an ARM Cortex-M4 microcontroller under deep learning(DL)-based side-channel attacks (DL-SCA), revealing that a single power trace enables 32-byte message recovery with 99.98% accuracy, while cross-device testing across three chips achieves over 99% accuracy, highlighting critical DL-driven threats. To better understand the spectral characteristics exploited by deep-learning-based side-channel attacks, we conduct a controlled frequency-band sensitivity study using Butterworth and Chebyshev low-pass filters applied to recorded traces. Across different cutoff frequencies and filter orders, the validation accuracy decreases substantially under aggressive attenuation, indicating that a significant fraction of exploitable leakage resides in higher-frequency components. Rather than claiming that offline digital filtering constrains a realistic adversary, we interpret these results as evidence for where leakage-sensitive information is concentrated, thereby providing guidance for implementable countermeasures such as masking, shuffling, desynchronization, and acquisition-path power/EM conditioning in embedded PQC systems. This cost-efficient solution requires no algorithm/hardware modifications, emphasizing the urgency of implementation-level protections for secure PQC deployment in resource-constrained systems.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Deep LearningComputer SecurityAlgorithms

Autor:innen und Institutionen

  • Department of Information Engineering and Computer Science, Feng Chia University, Taichung, Taiwan.
  • Master's Program of Information and Communication Security, Feng Chia University, Taichung, Taiwan.

ORCID-Profile: Chung-Wei Kuo

Förderung

  • National Science and Technology Council

Zitieren

APA

Kuo, C. W., Hung, Y. I., & Lin, K. L. (2026). Deep-learning side-channel analysis of embedded ML-KEM: Single-trace message recovery, cross-device generalization, and frequency-band leakage characterization. PLOS One, 21(11), e0358974. https://doi.org/10.1371/journal.pone.0358974

Vancouver

Kuo CW, Hung YI, Lin KL. Deep-learning side-channel analysis of embedded ML-KEM: Single-trace message recovery, cross-device generalization, and frequency-band leakage characterization. PLOS One. 2026;21(11):e0358974. doi:10.1371/journal.pone.0358974

BibTeX
@article{Kuo2026,
  author = {Kuo, Chung-Wei and Hung, Yu-I and Lin, Kai-Lun},
  title = {Deep-learning side-channel analysis of embedded ML-KEM: Single-trace message recovery, cross-device generalization, and frequency-band leakage characterization},
  journal = {PLOS One},
  year = {2026},
  volume = {21},
  number = {11},
  pages = {e0358974},
  doi = {10.1371/journal.pone.0358974}
}
Kostenlos starten →