Hacking Neural Networks: A Short Introduction

Michael Kissner
11/18/2019
cs.CRcs.LG

Official code: Kayzaks/HackingNeuralNetworks (2,641 stars, MIT, last commit 2020-02-29).

Abstract

A large chunk of research on the security issues of neural networks is focused on adversarial attacks. However, there exists a vast sea of simpler attacks one can perform both against and with neural networks. In this article, we give a quick introduction on how deep learning in security works and explore the basic methods of exploitation, but also look at the offensive capabilities deep learning enabled tools provide. All presented attacks, such as backdooring, GPU-based buffer overflows or automated bug hunting, are accompanied by short open-source exercises for anyone to try out.

View on arXivView PDF

Code Implementations(1)

Kayzaks/HackingNeuralNetworksOfficial100%
MIT

A small course on exploiting and defending neural networks

2,641306PythonOct 15, 20196 years agoMIT

Similar Papers

BackdoorBench: A Comprehensive Benchmark and Analysis of Backdoor Learning

88% match2024

A4O: All Trigger for One sample

88% match2025

FreeEagle: Detecting Complex Neural Trojans in Data-Free Cases

88% match1 repo202023

DeepBaR: Fault Backdoor Attack on Deep Neural Network Layers

87% match2024

Efficient Backdoor Attacks for Deep Neural Networks in Real-world Scenarios

87% match1 repo52023

AFLOW: Developing Adversarial Examples under Extremely Noise-limited Settings

87% match2023

Cite this paper

@article{kissner2019hacking,
  title  = {Hacking Neural Networks: A Short Introduction},
  author = {Michael Kissner},
  year   = {2019},
  eprint = {1911.07658},
  archivePrefix = {arXiv},
  url    = {https://arxiv.org/abs/1911.07658}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/1911.07658.svg)](https://codeofpaper.com/paper/1911.07658)

Discussion