Cinofix/sponge_poisoning_energy_latency_attack
Source code for the Energy-Latency Attacks via Sponge Poisoning paper.
This project helps evaluate the vulnerability of deep neural networks (DNNs) to 'sponge poisoning' attacks, which aim to increase a model's computational resource consumption (energy, latency) without degrading its accuracy. It takes a trained DNN model and a dataset as input and produces an 'attacked' version of the model, along with statistics and visualizations showing the increased energy consumption and latency. This tool is for AI/ML researchers, security analysts, and engineers evaluating the robustness and deployment costs of DNNs.
No commits in the last 6 months.
Use this if you need to understand how a deep learning model's resource consumption can be maliciously inflated while maintaining its predictive performance.
Not ideal if you are looking for methods to improve model efficiency or reduce inference costs through benign optimization techniques.
Stars
15
Forks
2
Language
Python
License
—
Category
Last pushed
Mar 14, 2022
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/Cinofix/sponge_poisoning_energy_latency_attack"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
QData/TextAttack
TextAttack 🐙 is a Python framework for adversarial attacks, data augmentation, and model...
ebagdasa/backdoors101
Backdoors Framework for Deep Learning and Federated Learning. A light-weight tool to conduct...
THUYimingLi/backdoor-learning-resources
A list of backdoor learning resources
zhangzp9970/MIA
Unofficial pytorch implementation of paper: Model Inversion Attacks that Exploit Confidence...
LukasStruppek/Plug-and-Play-Attacks
[ICML 2022 / ICLR 2024] Source code for our papers "Plug & Play Attacks: Towards Robust and...