dessertlab/Targeted-Data-Poisoning-Attacks

This repository contains the code, the dataset and the experimental results related to the paper "Vulnerabilities in AI Code Generators: Exploring Targeted Data Poisoning Attacks" accepted for publication at The 32nd IEEE/ACM International Conference on Program Comprehension (ICPC 2024).

32
/ 100
Emerging

This project explores how to inject software vulnerabilities into the training data of AI code generators. It takes existing datasets of safe code descriptions and snippets, modifies them to include vulnerable patterns, and then shows how this poisoned data affects AI models like CodeBERT. Security researchers and AI model developers would use this to understand and mitigate security risks in AI-generated code.

No commits in the last 6 months.

Use this if you are a security researcher or AI model developer investigating the security implications of AI code generators and want to reproduce targeted data poisoning attacks.

Not ideal if you are looking for a tool to run code generation tasks or for a general-purpose security scanner for AI-generated code.

AI-security software-vulnerability AI-model-auditing code-generation data-poisoning
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 5 / 25
Maturity 16 / 25
Community 11 / 25

How are scores calculated?

Stars

13

Forks

2

Language

Python

License

GPL-3.0

Last pushed

Aug 05, 2024

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/generative-ai/dessertlab/Targeted-Data-Poisoning-Attacks"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.