# Datasets

## Downloading the dataset

There will be only one public datasets given to train the neural networks.

The dataset can be downloaded from the following [link](https://drive.google.com/drive/folders/1JGbphwZXQvN_tEhpBIbQ-q-pN9wkqKQ-?usp=sharing).

It consists of 500K profiling traces and 100k attack traces. 

The hash of the above dataset is as follows:

```
132ae2e9a8213c983bf3b63449e9572d5d71d3b376a75b236415d4a728b9379f
```

There will be three private set of attack traces used for evaluation (see [Challenge Rule](https://pace-tl.gitbook.io/ches-challenge-2025/quickstart-4)).

The private dataset can be downloaded from the following [link](https://drive.google.com/drive/folders/1foaYhR_40_s17vmOdEv0dhKFVB09qksc?usp=drive_link).\
The hash of the private datasets (into one file) is as follow: 

```
86ec5b8fefb6ff9aad88112782ea34b6f3785bf9533c94a9ece51ccf9587ca97
```

## Flow of the public dataset

The datasets are stored in `.h5` file. The flow of the public dataset is similar to the ASCAD dataset \[1].

The structure of the public dataset is outlined as follows:

<figure><img src="https://3483517490-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Fv9irgr9f93ieYI6dbkvg%2Fuploads%2FleflrVggECL2qxKBbKTE%2Fh5_stuff.svg?alt=media&#x26;token=8ba28566-fc87-4fc3-97bd-21e3cc5ffd19" alt=""><figcaption><p>Dataset</p></figcaption></figure>

*Note: We will be targeting only the first byte for this challenge (i.e. byte* $$0$$*).*

## Flow of the private dataset

The structure of the private dataset is outlined as follows:

<figure><img src="https://3483517490-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Fv9irgr9f93ieYI6dbkvg%2Fuploads%2FRQb0xohsavll0hLwZPS8%2FCHES_Challenge_evaluation.svg?alt=media&#x26;token=7d1a069c-f320-42bb-b988-56542c5fefc2" alt=""><figcaption></figcaption></figure>

## Loading the datasets

The function to load the public dataset can be found in `utils.py`  as the function `load_ctf_2025()` . This function is used within `main_{tf/pytorch}.py` and `analyze_{tf/pytorch}.py`  (see [Getting Started](https://pace-tl.gitbook.io/ches-challenge-2025/quickstart)).

The private dataset can be loaded using the `load_data.py`  script available at the provided [link](https://drive.google.com/drive/folders/1foaYhR_40_s17vmOdEv0dhKFVB09qksc?usp=sharing). A `README.md` file is also included with detailed instructions on how to load the dataset.

## Citation

If our dataset contributed to your research, please acknowledge it with the following citation:

```
@misc{ge_wars2025,
author = {Shivam Bhasin and Harishma Boyapally and Dirmanto Jap and Trevor Yap and Qianmei Wu.},
title = {{GE Wars: The Deep Learning SCA battle}},
howpublished    = {CHES Challenge 2025},
year      = {2025},
note = {https://pace-tl.gitbook.io/ches-challenge-2025}
}
```

## References

1. Benadjila, R., Prouff, E., Strullu, R. *et al.* Deep learning for side-channel analysis and introduction to ASCAD database. *J Cryptogr Eng* 10, 163–188 (2020). <https://doi.org/10.1007/s13389-019-00220-8>


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://pace-tl.gitbook.io/ches-challenge-2025/quickstart-2.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.