pith. sign in

arxiv: 2605.22709 · v1 · pith:LQOEWD6Hnew · submitted 2026-05-21 · 💻 cs.CR · cs.ET· cs.RO· cs.SY· eess.SY

TriSweep: A Four-Drone Swarm Framework for Electromagnetic Side-Channel Analysis

Eric Yocam , Varghese Vaidyan This is my paper

Pith reviewed 2026-05-22 04:32 UTC · model grok-4.3

classification 💻 cs.CR cs.ETcs.ROcs.SYeess.SY
keywords electromagnetic side-channel analysisdrone swarmmasked AEScoherent combiningstandoff attackASCAD datasetkey recoverymask cancellation
0
0 comments X

The pith

A four-drone swarm extracts AES keys from masked microcontrollers at 0.25 meter standoff in simulation.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper establishes that a swarm of four drones can carry out electromagnetic side-channel analysis on embedded devices from distances of 0.25 to 1.5 meters by assigning specialized roles to three flying collector drones and routing their data to a stationary accumulator drone. The accumulator improves signal quality through coherent addition of the streams and removes second-order masking by computing a centered product across the spatially separated leakage traces. This setup is evaluated on real ANSSI ASCAD datasets of masked AES-128, including versions with timing desynchronization, and yields a simulated key rank of 18 plus or minus 1.7 at the closest range. Sympathetic readers would care because the work broadens the standard threat model of stationary close-range probes to include autonomous aerial platforms that could operate without physical access.

Core claim

TriSweep is a simulation framework that designs a four-drone swarm for standoff electromagnetic side-channel analysis. Three collector drones (Anchor for full-spectrum capture, Mask Probe for mask-register leakage, and Cipher Probe for masked SubBytes leakage) feed traces to a stationary Accumulator drone. The Accumulator applies coherent combining for a 4.8 dB SNR gain and performs second-order mask cancellation via the centered product of the two spatially separated streams. On the primary masked ASCAD dataset the framework reaches a simulated key rank of 18 plus or minus 1.7 at 0.25 m standoff distance, with cross-correlation alignment and a two-channel CNN further improving results on 50

What carries the argument

Four-drone swarm architecture with three spatially specialized collector drones feeding a stationary accumulator that performs coherent combining and centered-product mask cancellation.

If this is right

  • Coherent combining of traces from spatially separated drones yields a 4.8 dB SNR improvement.
  • Centered product of leakage streams from two different drones cancels second-order masking.
  • Cross-correlation alignment of profiling traces compensates for hover-induced jitter and reduces key rank on desynchronized data.
  • A two-channel CNN running in the accumulator lowers loss and improves rank on jittered datasets.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • If the simulation holds in hardware, security evaluations for embedded devices must incorporate multi-point aerial attack scenarios rather than single stationary probes.
  • The spatial separation technique could be tested on other leakage types such as power consumption or acoustic emissions to see whether similar mask cancellation occurs.
  • Real deployment would introduce new variables such as regulatory flight restrictions and drone detection that the current simulation leaves unaddressed.
  • Prototype testing could quantify additional noise sources from actual drone motors and wind that the propagation model may omit.

Load-bearing premise

The simulation of drone hover vibration, electromagnetic propagation over 0.25 to 1.5 meters, and the gains from coherent combining plus centered-product cancellation accurately represents physical reality.

What would settle it

Build and fly the physical four-drone prototype, record electromagnetic traces from a real masked AES microcontroller at 0.25 m standoff, and measure whether the recovered key rank reaches approximately 18.

Figures

Figures reproduced from arXiv: 2605.22709 by Eric Yocam, Varghese Vaidyan.

Figure 1
Figure 1. Figure 1: TriSweep four-drone system context. TriSweep comprises four drone platforms. Drones A, B, and C collect EM traces and forward synchronized IQ buffers to Drone D over 5 GHz Wi-Fi. Drone D (Accumulator) 6 [PITH_FULL_IMAGE:figures/full_fig_p006_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: TriSweep four-drone spatial layout. 7 [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: TriSweep end-to-end attack pipeline with all algorithm references. 5.1 Datasets ASCAD Masked (primary): 50,000 profiling + 10,000 attack traces (700 samples, AT￾mega8515, first-order masked AES-128). Baseline SNR = −22.9 dB [19]. Stored labels are masked; mask bytes are XOR-applied to recover unmasked labels for CNN training. ASCAD Desync-50/100: Same hardware with ±50 and ±100 sample random circular shift… view at source ↗
Figure 4
Figure 4. Figure 4: EM leakage spectrum, real ASCAD ATmega8515 masked AES-128 traces. [PITH_FULL_IMAGE:figures/full_fig_p015_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Simulated SNR vs. standoff distance for 1, 3, and 5 drones. [PITH_FULL_IMAGE:figures/full_fig_p015_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Four-drone ablation by drone count across all five datasets. [PITH_FULL_IMAGE:figures/full_fig_p016_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Multi-seed statistical validation on ASCAD_Masked ( [PITH_FULL_IMAGE:figures/full_fig_p016_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Key rank vs. trace count at five standoff distances, full four-drone system. [PITH_FULL_IMAGE:figures/full_fig_p017_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Key rank vs. trace count, three-drone CNN baseline. [PITH_FULL_IMAGE:figures/full_fig_p017_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Key rank vs. traces for 1, 2, and 3 drones at 1.0 m standoff. [PITH_FULL_IMAGE:figures/full_fig_p018_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: uses heterogeneous templates: Drone A from ASCAD_Masked, Drone B from ASCAD_Desync50, Drone C from synthetic fallback. Four-drone result (rank 92) is substan￾tially worse than the homogeneous case (rank 20), confirming matched profiling templates are required for effective second-order cancellation [PITH_FULL_IMAGE:figures/full_fig_p018_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Manual B×C second-order vs. CNN-enhanced Drone D, five-seed on AS￾CAD_Masked. 7 Discussion This section examines the TriSweep framework from four perspectives: design trade-offs inherent in the four-drone architecture, operational and environmental factors that will affect 19 [PITH_FULL_IMAGE:figures/full_fig_p019_12.png] view at source ↗
read the original abstract

Electromagnetic (EM) side-channel analysis traditionally assumes a stationary, close-proximity probe - a threat model that underestimates aerial adversaries. TriSweep is a simulation framework that designs and evaluates a four-drone swarm architecture for autonomous standoff EM-SCA of embedded microcontrollers at 0.25-1.5 m. Three spatially specialized collector drones - Anchor (full-spectrum), Mask Probe (mask-register loading leakage), and Cipher Probe (masked SubBytes output leakage) - feed a stationary Accumulator drone that performs coherent combining (+4.8 dB SNR gain) and second-order mask cancellation via a centered product of the two spatially separated leakage streams. Evaluated against three real ANSSI ASCAD datasets (ATmega8515 masked AES-128 and 50/100-sample desynchronized variants), the framework achieves a simulated key rank of 18 +/- 1.7 (five-seed) at 0.25 m on the primary masked dataset. Profiling-trace cross-correlation alignment reduces single-drone rank from 89 to 21 on the 100-sample-jitter variant, demonstrating compensation for drone hover vibration. A two-channel CNN in the Accumulator converges to a loss of 0.454 (vs. random baseline 5.545) and improves rank on desynchronized datasets. No physical hardware has been fabricated; prototype construction is the planned next step.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The manuscript introduces TriSweep, a simulation framework for a four-drone swarm performing standoff electromagnetic side-channel analysis on masked AES-128 implementations using real ANSSI ASCAD datasets. Three collector drones (Anchor, Mask Probe, Cipher Probe) capture spatially specialized leakage streams that feed a stationary Accumulator drone applying coherent combining (+4.8 dB SNR gain) and centered-product second-order mask cancellation. The framework reports a simulated key rank of 18 ± 1.7 (five seeds) at 0.25 m on the primary masked dataset, with cross-correlation alignment reducing rank from 89 to 21 on the 100-sample desynchronized variant and a two-channel CNN converging to loss 0.454 on jittered traces. No physical hardware prototype is presented.

Significance. If the modeled standoff propagation, hover vibration compensation, and coherent combining accurately reflect physical behavior, the work would establish a credible new threat model for aerial EM-SCA adversaries and demonstrate how spatial separation can enable mask cancellation without explicit knowledge of the mask. The reuse of established ASCAD traces together with explicit SNR and rank metrics provides a reproducible baseline for exploring mobile side-channel attacks.

major comments (2)
  1. [Abstract] Abstract and Evaluation: The headline result (key rank 18 ± 1.7 at 0.25 m with +4.8 dB SNR gain and centered-product mask cancellation) is obtained by applying standard SCA combining operations to external ASCAD traces inside a simulation; because no physical drone measurements or hardware prototype exist, unmodeled factors such as motor EMI, multipath, or imperfect spatial separation remain untested and directly affect whether the reported gains would appear in practice.
  2. [Evaluation] Evaluation section: The vibration-mitigation claim relies on profiling-trace cross-correlation alignment reducing single-drone rank from 89 to 21 on the 100-sample-jitter variant; the manuscript does not report the sensitivity of this alignment to the specific vibration amplitude or frequency model used, leaving open whether the compensation generalizes beyond the simulated conditions.
minor comments (2)
  1. [Abstract] Abstract: The framework is named TriSweep yet deploys four drones; a brief clarification of the naming rationale would avoid reader confusion.
  2. [Evaluation] The two-channel CNN loss of 0.454 versus random baseline 5.545 is reported without stating the exact loss function or the number of training epochs, which would aid reproducibility.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback on our simulation framework for a four-drone EM side-channel swarm. We address each major comment below, clarifying the simulation scope while enhancing the manuscript to better highlight limitations and evaluation robustness. The work establishes a reproducible baseline using real ASCAD traces; physical validation remains planned future work.

read point-by-point responses

  1. Referee: [Abstract] Abstract and Evaluation: The headline result (key rank 18 ± 1.7 at 0.25 m with +4.8 dB SNR gain and centered-product mask cancellation) is obtained by applying standard SCA combining operations to external ASCAD traces inside a simulation; because no physical drone measurements or hardware prototype exist, unmodeled factors such as motor EMI, multipath, or imperfect spatial separation remain untested and directly affect whether the reported gains would appear in practice.

    Authors: We agree that all results derive from simulation applying standard SCA operations to the ANSSI ASCAD datasets rather than new physical drone captures. The manuscript contribution is the swarm architecture design, spatial specialization of collectors, and coherent combining plus centered-product mask cancellation within a standoff propagation and vibration model. We have revised the abstract to emphasize the simulation nature and added a dedicated Limitations section discussing unmodeled effects including motor EMI, multipath, and spatial separation imperfections. These factors are acknowledged as requiring hardware validation, which is stated as the planned next step. The reported +4.8 dB SNR gain follows directly from the coherent summation model at the modeled standoff distance. revision: partial

  2. Referee: [Evaluation] Evaluation section: The vibration-mitigation claim relies on profiling-trace cross-correlation alignment reducing single-drone rank from 89 to 21 on the 100-sample-jitter variant; the manuscript does not report the sensitivity of this alignment to the specific vibration amplitude or frequency model used, leaving open whether the compensation generalizes beyond the simulated conditions.

    Authors: The cross-correlation alignment compensates for the 100-sample jitter variant that models drone hover vibration. We have added a sensitivity study in the revised Evaluation section that varies jitter amplitude from 50 to 150 samples while keeping the frequency characteristics fixed to the original model. The alignment continues to reduce key rank across this range (from 89 down to 18-25), supporting generalization within the simulated vibration parameters drawn from typical drone literature. We have also clarified the exact jitter distribution and alignment window used. revision: yes

Circularity Check

0 steps flagged

No significant circularity: results derive from standard SCA processing on external ASCAD datasets

full rationale

The paper's central results (key rank 18 +/- 1.7 at 0.25 m, +4.8 dB SNR gain, mask cancellation) are obtained by applying established techniques—coherent combining, centered-product second-order cancellation, cross-correlation alignment, and a two-channel CNN—to real, publicly available ANSSI ASCAD traces inside a simulation of drone effects. These operations are not fitted to the target metric and then re-predicted; the simulation merely adds modeled propagation and vibration effects to independent external data. No self-citation chains, uniqueness theorems, or ansatzes imported from prior author work are load-bearing. The derivation chain remains self-contained against the external benchmarks.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The framework rests on standard EM leakage models and combining techniques from prior SCA literature plus unvalidated assumptions about drone positioning accuracy and signal propagation at standoff distances.

free parameters (1)
  • Coherent combining SNR gain
    The reported +4.8 dB gain is presented as a simulation outcome but depends on modeled drone positions and channel assumptions that are not independently derived.
axioms (1)
  • domain assumption EM leakage from masked AES can be spatially separated and combined to cancel first-order masking
    Invoked when describing the Mask Probe and Cipher Probe roles and the centered-product cancellation in the Accumulator.

pith-pipeline@v0.9.0 · 5792 in / 1279 out tokens · 38495 ms · 2026-05-22T04:32:15.097507+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

44 extracted references · 44 canonical work pages

  1. [1]

    Electromagnetic analysis: Concrete results,

    K. Gandolfi, C. Mourtel, and F. Olivier, “Electromagnetic analysis: Concrete results,” inCryptographic Hardware and Embedded Systems — CHES 2001, ser. Lecture Notes in Computer Science, vol. 2162. Berlin, Heidelberg: Springer, 2001, pp. 251–261

    work page 2001

  2. [2]

    The EM side-channel(s),

    D. Agrawal, B. Archambeault, J. R. Rao, and P. Rohatgi, “The EM side-channel(s),” in Cryptographic Hardware and Embedded Systems — CHES 2002, ser. Lecture Notes in Computer Science, vol. 2523. Berlin, Heidelberg: Springer, 2002, pp. 29–45

    work page 2002

  3. [3]

    Differentialpoweranalysis,

    P.C.Kocher, J.Jaffe, andB.Jun, “Differentialpoweranalysis,” inAdvances in Cryptology — CRYPTO ’99, ser. Lecture Notes in Computer Science, vol. 1666. Berlin, Heidelberg: Springer, 1999, pp. 388–397

    work page 1999

  4. [4]

    Mangard, E

    S. Mangard, E. Oswald, and T. Popp,Power Analysis Attacks: Revealing the Secrets of Smart Cards. New York: Springer, 2007

    work page 2007

  5. [5]

    Correlation power analysis with a leakage model,

    E. Brier, C. Clavier, and F. Olivier, “Correlation power analysis with a leakage model,” inCryptographic Hardware and Embedded Systems — CHES 2004, ser. Lecture Notes in Computer Science, vol. 3156. Berlin, Heidelberg: Springer, 2004, pp. 16–29

    work page 2004

  6. [6]

    Template attacks,

    S. Chari, J. R. Rao, and P. Rohatgi, “Template attacks,” inCryptographic Hardware and Embedded Systems — CHES 2002, ser. Lecture Notes in Computer Science, vol

    work page 2002

  7. [7]

    Berlin, Heidelberg: Springer, 2002, pp. 13–28

    work page 2002

  8. [8]

    Using second-order power analysis to attack DPA resistant software,

    T. S. Messerges, “Using second-order power analysis to attack DPA resistant software,” inCryptographic Hardware and Embedded Systems — CHES 2000, ser. Lecture Notes in Computer Science, vol. 1965. Berlin, Heidelberg: Springer, 2000, pp. 238–251

    work page 2000

  9. [9]

    Masking against side-channel attacks: A formal security proof,

    E. Prouff and M. Rivain, “Masking against side-channel attacks: A formal security proof,” inAdvances in Cryptology — EUROCRYPT 2013, ser. Lecture Notes in Computer Science, vol. 7881. Berlin, Heidelberg: Springer, 2013, pp. 142–159

    work page 2013

  10. [10]

    Breaking cryptographic implementations using deep learning techniques,

    H. Maghrebi, T. Portigliatti, and E. Prouff, “Breaking cryptographic implementations using deep learning techniques,” inSecurity, Privacy, and Applied Cryptography Engi- neering (SPACE 2016), ser. Lecture Notes in Computer Science, vol. 10076. Cham: Springer, 2016, pp. 3–26

    work page 2016

  11. [11]

    Methodology for efficient CNN architectures in profiling attacks,

    G. Zaid, L. Bossuet, A. Habrard, and A. Venelli, “Methodology for efficient CNN architectures in profiling attacks,”IACR Transactions on Cryptographic Hardware and Embedded Systems (TCHES), vol. 2020, no. 1, pp. 1–36, 2020

    work page 2020

  12. [12]

    Advanced encryption standard (AES),

    National Institute of Standards and Technology, “Advanced encryption standard (AES),” NIST, FIPS Publication 197, 2001

    work page 2001

  13. [13]

    A tutorial on UAVs for wireless networks: Applications, challenges, and open problems,

    M. Mozaffari, W. Saad, M. Bennis, Y.-H. Nam, and M. Debbah, “A tutorial on UAVs for wireless networks: Applications, challenges, and open problems,”IEEE Communications Surveys and Tutorials, vol. 21, no. 3, pp. 2334–2360, 2019

    work page 2019

  14. [14]

    The software radio architecture,

    J. Mitola, “The software radio architecture,”IEEE Communications Magazine, vol. 33, no. 5, pp. 26–38, 1995. 22

    work page 1995

  15. [15]

    STELLAR: A generic EM side-channel attack protection through ground-up root-cause analysis,

    D. Das, M. Nath, S. Ghosh, A. Raychowdhury, and S. Sen, “STELLAR: A generic EM side-channel attack protection through ground-up root-cause analysis,” in2019 IEEE International Symposium on Hardware Oriented Security and Trust (HOST). IEEE, 2019, pp. 11–20

    work page 2019

  16. [16]

    The vulnerability of UAVs to cyber attacks — an approach to the risk assessment,

    K. Hartmann and C. Steup, “The vulnerability of UAVs to cyber attacks — an approach to the risk assessment,” in5th International Conference on Cyber Conflict (CyCon 2013). Tallinn, Estonia: NATO CCD COE Publications, 2013, pp. 1–23. [Online]. Available: https://ieeexplore.ieee.org/document/6569555

    work page arXiv 2013

  17. [17]

    Robotics cyber security: Vulnerabilities, attacks, countermeasures, and recommendations,

    J.-P. A. Yaacoub, H. N. Noura, O. Salman, and A. Chehab, “Robotics cyber security: Vulnerabilities, attacks, countermeasures, and recommendations,”International Journal of Information Security, vol. 21, no. 1, pp. 115–158, 2022

    work page 2022

  18. [18]

    H. L. Van Trees,Optimum Array Processing: Part IV of Detection, Estimation, and Modulation Theory. New York: Wiley-Interscience, 2002

    work page 2002

  19. [19]

    Breaking masked implementations with many shares on 32-bit software platforms,

    O. Bronchain and F.-X. Standaert, “Breaking masked implementations with many shares on 32-bit software platforms,”IACR Transactions on Cryptographic Hardware and Embedded Systems (TCHES), vol. 2021, no. 3, pp. 202–234, 2021

    work page 2021

  20. [20]

    Deep learning for side- channel analysis and introduction to ASCAD database,

    R. Benadjila, E. Prouff, R. Strullu, E. Cagli, and C. Dumas, “Deep learning for side- channel analysis and introduction to ASCAD database,”Journal of Cryptographic Engineering, vol. 10, no. 2, pp. 163–188, 2020

    work page 2020

  21. [21]

    ElectroMagnetic analysis (EMA): Measures and counter-measures for smart cards,

    J.-J. Quisquater and D. Samyde, “ElectroMagnetic analysis (EMA): Measures and counter-measures for smart cards,” inSmart Card Programming and Security (e- Smart 2001), ser. Lecture Notes in Computer Science, vol. 2140. Berlin, Heidelberg: Springer, 2001, pp. 200–210

    work page 2001

  22. [22]

    Examining smart-card security under the threat of power analysis attacks,

    T. S. Messerges, E. A. Dabbish, and R. H. Sloan, “Examining smart-card security under the threat of power analysis attacks,”IEEE Transactions on Computers, vol. 51, no. 5, pp. 541–552, 2002

    work page 2002

  23. [23]

    Localized electromagnetic analysis of cryptographic implementations,

    J. Heyszl, S. Mangard, B. Heinz, F. Stumpf, and G. Sigl, “Localized electromagnetic analysis of cryptographic implementations,” inTopics in Cryptology — CT-RSA 2012, ser. Lecture Notes in Computer Science, vol. 7178. Berlin, Heidelberg: Springer, 2012, pp. 231–244

    work page 2012

  24. [24]

    Fault injection attacks on cryptographic devices: Theory, practice, and countermeasures,

    A. Barenghi, L. Breveglieri, I. Koren, and D. Naccache, “Fault injection attacks on cryptographic devices: Theory, practice, and countermeasures,”Proceedings of the IEEE, vol. 100, no. 11, pp. 3056–3076, 2012

    work page 2012

  25. [25]

    PLATYPUS: Software-based power side-channel attacks on x86,

    M. Lipp, A. Kogler, D. A. Oswald, M. Schwarz, C. Easdon, C. Canella, and D. Gruss, “PLATYPUS: Software-based power side-channel attacks on x86,” in2021 IEEE Sympo- sium on Security and Privacy (SP). IEEE, 2021, pp. 355–371

    work page 2021

  26. [26]

    On second-order differential power analysis,

    M. Joye, P. Paillier, and B. Schoenmakers, “On second-order differential power analysis,” inCryptographic Hardware and Embedded Systems — CHES 2003, ser. Lecture Notes in Computer Science, vol. 2779. Berlin, Heidelberg: Springer, 2003, pp. 293–308

    work page 2003

  27. [27]

    Statistical analysis of second order differential power analysis,

    E. Prouff, M. Rivain, and R. Bevan, “Statistical analysis of second order differential power analysis,”IEEE Transactions on Computers, vol. 58, no. 6, pp. 799–811, 2009. 23

    work page 2009

  28. [28]

    Towards efficient second-order power analysis,

    J. Waddle and D. Wagner, “Towards efficient second-order power analysis,” inCrypto- graphic Hardware and Embedded Systems — CHES 2004, ser. Lecture Notes in Computer Science, vol. 3156. Berlin, Heidelberg: Springer, 2004, pp. 1–15

    work page 2004

  29. [29]

    Convolutional neural networks with data augmen- tation against jitter-based countermeasures,

    E. Cagli, C. Dumas, and E. Prouff, “Convolutional neural networks with data augmen- tation against jitter-based countermeasures,” inCryptographic Hardware and Embedded Systems — CHES 2017, ser. Lecture Notes in Computer Science, vol. 10529. Cham: Springer, 2017, pp. 45–68

    work page 2017

  30. [30]

    Make some noise: Unleashing the power of convolutional neural networks for profiled side-channel analysis,

    J. Kim, S. Picek, A. Heuser, S. Bhasin, and A. Hanjalic, “Make some noise: Unleashing the power of convolutional neural networks for profiled side-channel analysis,”IACR Transactions on Cryptographic Hardware and Embedded Systems (TCHES), vol. 2019, no. 3, pp. 148–179, 2019

    work page 2019

  31. [31]

    Revisiting a methodology for efficient CNN architectures in profiling attacks,

    L. Wouters, B. Gierlichs, and B. Preneel, “Revisiting a methodology for efficient CNN architectures in profiling attacks,”IACR Transactions on Cryptographic Hardware and Embedded Systems (TCHES), vol. 2022, no. 1, pp. 147–168, 2022

    work page 2022

  32. [32]

    Strength in numbers: Improving generalization with ensembles in machine learning-based profiled side-channel analysis,

    G. Perin, Ł. Chmielewski, and S. Picek, “Strength in numbers: Improving generalization with ensembles in machine learning-based profiled side-channel analysis,”IACR Trans- actions on Cryptographic Hardware and Embedded Systems (TCHES), vol. 2020, no. 4, pp. 337–364, 2020

    work page 2020

  33. [33]

    I choose you: Automated hyperparameter tuning for deep learning-based side-channel analysis,

    L. Wu, G. Perin, and S. Picek, “I choose you: Automated hyperparameter tuning for deep learning-based side-channel analysis,”IACR Transactions on Cryptographic Hardware and Embedded Systems (TCHES), vol. 2022, no. 3, pp. 325–353, 2022

    work page 2022

  34. [34]

    The curse of class imbalance and conflicting metrics with machine learning for side-channel evaluations,

    S. Picek, A. Heuser, A. Jovic, S. Bhasin, and F. Regazzoni, “The curse of class imbalance and conflicting metrics with machine learning for side-channel evaluations,”IACR Transactions on Cryptographic Hardware and Embedded Systems (TCHES), vol. 2019, no. 1, pp. 209–237, 2019

    work page 2019

  35. [35]

    SoK: Deep learning-based physical side-channel analysis,

    S. Picek, G. Perin, L. Mariot, L. Wu, and L. Batina, “SoK: Deep learning-based physical side-channel analysis,”ACM Computing Surveys, vol. 55, no. 11, pp. 236:1–236:35, 2023

    work page 2023

  36. [36]

    Compromising electromagnetic emanations of wired and wireless keyboards,

    M. Vuagnoux and S. Pasini, “Compromising electromagnetic emanations of wired and wireless keyboards,” inProceedings of the 18th USENIX Security Symposium. Berkeley, CA, USA: USENIX Association, 2009, pp. 1–16. [Online]. Available: https://www.usenix.org/legacy/events/sec09/tech/full_papers/vuagnoux.pdf

    work page 2009

  37. [37]

    Get your hands off my laptop: Physical side- channel key-extraction attacks on PCs,

    D. Genkin, I. Pipman, and E. Tromer, “Get your hands off my laptop: Physical side- channel key-extraction attacks on PCs,” inCryptographic Hardware and Embedded Systems — CHES 2015, ser. Lecture Notes in Computer Science, vol. 9293. Berlin, Heidelberg: Springer, 2015, pp. 130–150

    work page 2015

  38. [38]

    Screaming channels: When electromagnetic side channels meet radio transceivers,

    G. Camurati, S. Poeplau, M. Muench, T. Hayes, and A. Francillon, “Screaming channels: When electromagnetic side channels meet radio transceivers,” inProceedings of the 2018 ACM SIGSAC Conference on Computer and Communications Security (CCS). New York, NY, USA: ACM, 2018, pp. 163–177

    work page 2018

  39. [39]

    On configurable SCA coun- termeasures against single trace attacks for the NTT: A performance evaluation study 24 over KYBER and DILITHIUM on the ARM Cortex-M4,

    P. Ravi, R. Poussier, S. Bhasin, and A. Chattopadhyay, “On configurable SCA coun- termeasures against single trace attacks for the NTT: A performance evaluation study 24 over KYBER and DILITHIUM on the ARM Cortex-M4,”IACR Transactions on Cryp- tographic Hardware and Embedded Systems (TCHES), vol. 2022, no. 3, pp. 156–192, 2022

    work page 2022

  40. [40]

    A survey on electromagnetic side- channel attacks and discussion on their case-detecting possibilities for digital forensics,

    A. P. Sayakkara, N.-A. Le-Khac, and M. Scanlon, “A survey on electromagnetic side- channel attacks and discussion on their case-detecting possibilities for digital forensics,” Digital Investigation, vol. 29, pp. 43–54, 2019

    work page 2019

  41. [41]

    Template attacks in principal subspaces,

    C. Archambeau, E. Peeters, F.-X. Standaert, and J.-J. Quisquater, “Template attacks in principal subspaces,” inCryptographic Hardware and Embedded Systems — CHES 2006, ser. Lecture Notes in Computer Science, vol. 4249. Berlin, Heidelberg: Springer, 2006, pp. 1–14

    work page 2006

  42. [42]

    A unified framework for the analysis of side-channel key recovery attacks,

    F.-X. Standaert, T. G. Malkin, and M. Yung, “A unified framework for the analysis of side-channel key recovery attacks,” inAdvances in Cryptology — EUROCRYPT 2009, ser. Lecture Notes in Computer Science, vol. 5479. Berlin, Heidelberg: Springer, 2009, pp. 443–461

    work page 2009

  43. [43]

    Goodfellow, Y

    I. Goodfellow, Y. Bengio, and A. Courville,Deep Learning. Cambridge, MA: MIT Press, 2016. [Online]. Available: http://www.deeplearningbook.org

    work page 2016

  44. [44]

    Leveling Dilithium against leakage: Revisiting sensitivity analysis on a state-of-the-art signature scheme,

    M. Azouaoui, O. Bronchain, C. Hoffmann, Y. Kuzovkova, T. Schneider, and F.-X. Standaert, “Leveling Dilithium against leakage: Revisiting sensitivity analysis on a state-of-the-art signature scheme,”IACR Transactions on Cryptographic Hardware and Embedded Systems (TCHES), vol. 2022, no. 4, pp. 674–703, 2022. 25

    work page 2022