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

We concentrate on machine learning techniques used for profiled side-channel analysis in the presence of imbalanced data. Such scenarios are realistic and often occurring, for instance in the Hamming weight or Hamming distance leakage models. In order to deal with the imbalanced data, we use various...

Full description

Saved in:
Bibliographic Details
Main Authors: Picek, Stjepan, Heuser, Annelie, Jovic, Alan, Bhasin, Shivam, Regazzoni, Francesco
Other Authors: Temasek Laboratories @ NTU
Format: Article
Language:English
Published: 2022
Subjects:
Online Access:https://hdl.handle.net/10356/160593
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
id sg-ntu-dr.10356-160593
record_format dspace
spelling sg-ntu-dr.10356-1605932022-07-30T20:12:32Z The curse of class imbalance and conflicting metrics with machine learning for side-channel evaluations Picek, Stjepan Heuser, Annelie Jovic, Alan Bhasin, Shivam Regazzoni, Francesco Temasek Laboratories @ NTU Engineering::Computer science and engineering Profiled Side-Channel Attacks Imbalanced Datasets We concentrate on machine learning techniques used for profiled side-channel analysis in the presence of imbalanced data. Such scenarios are realistic and often occurring, for instance in the Hamming weight or Hamming distance leakage models. In order to deal with the imbalanced data, we use various balancing techniques and we show that most of them help in mounting successful attacks when the data is highly imbalanced. Especially, the results with the SMOTE technique are encouraging, since we observe some scenarios where it reduces the number of necessary measurements more than 8 times. Next, we provide extensive results on comparison of machine learning and side-channel metrics, where we show that machine learning metrics (and especially accuracy as the most often used one) can be extremely deceptive. This finding opens a need to revisit the previous works and their results in order to properly assess the performance of machine learning in side-channel analysis. Published version 2022-07-27T05:46:15Z 2022-07-27T05:46:15Z 2019 Journal Article Picek, S., Heuser, A., Jovic, A., Bhasin, S. & Regazzoni, F. (2019). The curse of class imbalance and conflicting metrics with machine learning for side-channel evaluations. IACR Transactions On Cryptographic Hardware and Embedded Systems, 2019(1), 209-237. https://dx.doi.org/10.13154/tches.v2019.i1.209-237 2569-2925 https://hdl.handle.net/10356/160593 10.13154/tches.v2019.i1.209-237 2-s2.0-85118420091 1 2019 209 237 en IACR Transactions on Cryptographic Hardware and Embedded Systems © 2018 Stjepan Picek, Annelie Heuser, Alan Jovic, Shivam Bhasin, Francesco Regazzoni. This work is licensed under a Creative Commons Attribution 4.0 International License. application/pdf
institution Nanyang Technological University
building NTU Library
continent Asia
country Singapore
Singapore
content_provider NTU Library
collection DR-NTU
language English
topic Engineering::Computer science and engineering
Profiled Side-Channel Attacks
Imbalanced Datasets
spellingShingle Engineering::Computer science and engineering
Profiled Side-Channel Attacks
Imbalanced Datasets
Picek, Stjepan
Heuser, Annelie
Jovic, Alan
Bhasin, Shivam
Regazzoni, Francesco
The curse of class imbalance and conflicting metrics with machine learning for side-channel evaluations
description We concentrate on machine learning techniques used for profiled side-channel analysis in the presence of imbalanced data. Such scenarios are realistic and often occurring, for instance in the Hamming weight or Hamming distance leakage models. In order to deal with the imbalanced data, we use various balancing techniques and we show that most of them help in mounting successful attacks when the data is highly imbalanced. Especially, the results with the SMOTE technique are encouraging, since we observe some scenarios where it reduces the number of necessary measurements more than 8 times. Next, we provide extensive results on comparison of machine learning and side-channel metrics, where we show that machine learning metrics (and especially accuracy as the most often used one) can be extremely deceptive. This finding opens a need to revisit the previous works and their results in order to properly assess the performance of machine learning in side-channel analysis.
author2 Temasek Laboratories @ NTU
author_facet Temasek Laboratories @ NTU
Picek, Stjepan
Heuser, Annelie
Jovic, Alan
Bhasin, Shivam
Regazzoni, Francesco
format Article
author Picek, Stjepan
Heuser, Annelie
Jovic, Alan
Bhasin, Shivam
Regazzoni, Francesco
author_sort Picek, Stjepan
title The curse of class imbalance and conflicting metrics with machine learning for side-channel evaluations
title_short The curse of class imbalance and conflicting metrics with machine learning for side-channel evaluations
title_full The curse of class imbalance and conflicting metrics with machine learning for side-channel evaluations
title_fullStr The curse of class imbalance and conflicting metrics with machine learning for side-channel evaluations
title_full_unstemmed The curse of class imbalance and conflicting metrics with machine learning for side-channel evaluations
title_sort curse of class imbalance and conflicting metrics with machine learning for side-channel evaluations
publishDate 2022
url https://hdl.handle.net/10356/160593
_version_ 1739837471649169408