EstraNet: An Efficient Shift-Invariant Transformer Network for Side Channel Analysis

This repository contains the implementation of EstraNet, an efficient shift-invariant transformer network for Side-Channel Analysis.
For more details, refer to the paper.

---

Repository Structure

• fast_attention.py – Implements the proposed GaussiP attention layer.
• normalization.py – Implements the layer-centering normalization.
• transformer.py – Defines the EstraNet model architecture.
• train_trans.py – Training and evaluation script for EstraNet.
• data_utils.py – Utilities for loading ASCADf and ASCADr datasets.
• data_utils_ches20.py – Utilities for loading the CHES20 dataset.
• evaluation_utils.py – Computes guessing entropy for ASCAD datasets.
• evaluation_utils_ches20.py – Computes guessing entropy for CHES20 dataset.
• run_trans_\<dataset\>.sh – Bash scripts with predefined hyperparameters for specific datasets, where <dataset> is one of:
  • ASCADf (fixed key)
  • ASCADr (random key)
  • CHES20 (CHES CTF 2020)

---

Data Pre-processing:

• For the CHES CTF 2020 dataset, the traces are multiplied by a constant 0.004 to keep the feature value range within [-120, 120].

---

Tested on

• Python 3.8.10
• absl-py == 2.3.1
• numpy == 1.24.3
• scipy == 1.10.1
• h5py == 3.11.0
• tensorflow == 2.13.0

---

Getting Started

1. Clone the repository:

   git clone https://github.com/suvadeep-iitb/EstraNet.git
   cd EstraNet

2. Install dependencies (Python >= 3.8 recommended):

   pip install -r requirements.txt

3. Set dataset path in the bash script:

   Open run_trans_\<dataset\>.sh and set the dataset path variable properly.
   

4. Train EstraNet:

   bash run_trans_\<dataset\>.sh train

5. Perform Evaluation:

   bash run_trans_\<dataset\>.sh test

---

Citation:

@article{DBLP:journals/tches/HajraCM24,
  author       = {Suvadeep Hajra and
                  Siddhartha Chowdhury and
                  Debdeep Mukhopadhyay},
  title        = {EstraNet: An Efficient Shift-Invariant Transformer Network for Side-Channel
                  Analysis},
  journal      = {{IACR} Trans. Cryptogr. Hardw. Embed. Syst.},
  volume       = {2024},
  number       = {1},
  pages        = {336--374},
  year         = {2024},
  url          = {https://doi.org/10.46586/tches.v2024.i1.336-374},
  doi          = {10.46586/TCHES.V2024.I1.336-374},
  timestamp    = {Sat, 08 Jun 2024 13:14:59 +0200},
  biburl       = {https://dblp.org/rec/journals/tches/HajraCM24.bib},
  bibsource    = {dblp computer science bibliography, https://dblp.org}
}