Kayzaks/HackingNeuralNetworks
An educational course teaching methods to exploit and defend neural networks through hands-on exercises.
Not currently ranked — collecting fresh signals.
star history
This repository is a short introductory course on the security of neural networks. It covers offensive techniques such as bug hunting, shellcode obfuscation, and model exploitation methods like information extraction and backdooring, along with corresponding defensive strategies. The course provides exercises using Keras with TensorFlow backend, NumPy, SciPy, and other ML libraries, and is accompanied by a published article available on arXiv.
Frequently asked
- What is Kayzaks/HackingNeuralNetworks?
- An educational course teaching methods to exploit and defend neural networks through hands-on exercises.
- Is HackingNeuralNetworks open source?
- Yes — Kayzaks/HackingNeuralNetworks is open source, released under the MIT license.
- What language is HackingNeuralNetworks written in?
- Kayzaks/HackingNeuralNetworks is primarily written in Python.
- How popular is HackingNeuralNetworks?
- Kayzaks/HackingNeuralNetworks has 2.6k stars on GitHub.
- Where can I find HackingNeuralNetworks?
- Kayzaks/HackingNeuralNetworks is on GitHub at https://github.com/Kayzaks/HackingNeuralNetworks.