Published July 24, 2025 | Version v1
Journal Restricted

A Taxonomy of System-Level Attacks on Deep Learning Models in Autonomous Vehicles

  • 1. ROR icon Università della Svizzera italiana
  • 1. ROR icon University of Trento
  • 2. ROR icon Università della Svizzera italiana

Description

Replication Package for the Paper

"A Taxonomy of System-Level Attacks on Deep Learning Models in Autonomous Vehicles"

This repository contains the replication package associated with the paper titled "A Taxonomy of System-Level Attacks on Deep Learning Models in Autonomous Vehicles."

Contents

  • ChatGPT_for_AVs.pdf: This document presents the results of a structured ChatGPT search aimed at identifying the relevant domains of autonomous vehicles, deep learning models and modules, and simulation environments.

How to Replicate

  1. Install findpapers using the following command:

    pip install findpapers
    
  2. Open the send_query.py file to find the query used for paper extraction. Copy and execute the query to retrieve relevant research papers.

  3. Update the json_dir_path = './findpapers' variable in venue_filter.py to match the directory containing the list of downloaded papers.

  4. Execute venue_filter.py to filter papers based on the venues specified in the chosen_venues.txt file.

  5. The output will be a file named papers_after_venue_filter.csv, which includes the title, abstract, URL, venue, and publication date for each filtered paper.

Note: The venue_freq.py script calculates the frequency of papers per venue, distinguishing between journal and conference venues. Running this script is optional and was primarily used during our venue selection process.

Citation

Tehrani, M. J., Kim, J., Foulefack, R. Z. L., Marchetto, A., & Tonella, P. (2024).
A Taxonomy of System-Level Attacks on Deep Learning Models in Autonomous Vehicles.
https://arxiv.org/abs/2412.04510

@article{tehrani2024taxonomy,
  title={A Taxonomy of System-Level Attacks on Deep Learning Models in Autonomous Vehicles},
  author={Tehrani, Masoud Jamshidiyan and Kim, Jinhan and Foulefack, Rosmael Zidane Lekeufack and Marchetto, Alessandro and Tonella, Paolo},
  journal={arXiv preprint arXiv:2412.04510},
  year={2024}
}

 

Files

Restricted

The record is publicly accessible, but files are restricted to users with access.

Additional details

Funding

European Commission
Sec4AI4Sec - Cybersecurity for AI-Augmented Systems 101120393

Software