Deep neural networks are widely used for nonlinear function approximation
with applications ranging from computer vision to control. Although these
networks involve the composition of simple arithmetic operations, it can be
very challenging to verify whether a particular network satisfies certain
input-output properties. This article surveys methods that have emerged
recently for soundly verifying such properties. These methods borrow insights
from reachability analysis, optimization, and search. We discuss fundamental
differences and connections between existing algorithms. In addition, we
provide pedagogical implementations of existing methods and compare them on a
set of benchmark problems.
Description
[1903.06758] Algorithms for Verifying Deep Neural Networks
%0 Journal Article
%1 liu2019algorithms
%A Liu, Changliu
%A Arnon, Tomer
%A Lazarus, Christopher
%A Barrett, Clark
%A Kochenderfer, Mykel J.
%D 2019
%K deep-learning robustness
%T Algorithms for Verifying Deep Neural Networks
%U http://arxiv.org/abs/1903.06758
%X Deep neural networks are widely used for nonlinear function approximation
with applications ranging from computer vision to control. Although these
networks involve the composition of simple arithmetic operations, it can be
very challenging to verify whether a particular network satisfies certain
input-output properties. This article surveys methods that have emerged
recently for soundly verifying such properties. These methods borrow insights
from reachability analysis, optimization, and search. We discuss fundamental
differences and connections between existing algorithms. In addition, we
provide pedagogical implementations of existing methods and compare them on a
set of benchmark problems.
@article{liu2019algorithms,
abstract = {Deep neural networks are widely used for nonlinear function approximation
with applications ranging from computer vision to control. Although these
networks involve the composition of simple arithmetic operations, it can be
very challenging to verify whether a particular network satisfies certain
input-output properties. This article surveys methods that have emerged
recently for soundly verifying such properties. These methods borrow insights
from reachability analysis, optimization, and search. We discuss fundamental
differences and connections between existing algorithms. In addition, we
provide pedagogical implementations of existing methods and compare them on a
set of benchmark problems.},
added-at = {2019-03-22T13:05:32.000+0100},
author = {Liu, Changliu and Arnon, Tomer and Lazarus, Christopher and Barrett, Clark and Kochenderfer, Mykel J.},
biburl = {https://www.bibsonomy.org/bibtex/20673d1f5cc9f1562e9529c2c0d9a93d2/kirk86},
description = {[1903.06758] Algorithms for Verifying Deep Neural Networks},
interhash = {8e341c8be362cffd8fe42d21a2669aa5},
intrahash = {0673d1f5cc9f1562e9529c2c0d9a93d2},
keywords = {deep-learning robustness},
note = {cite arxiv:1903.06758},
timestamp = {2019-03-22T13:05:32.000+0100},
title = {Algorithms for Verifying Deep Neural Networks},
url = {http://arxiv.org/abs/1903.06758},
year = 2019
}