TY - JOUR
T1 - Exponent Relaxation of Polynomial Zonotopes and Its Applications in Formal Neural Network Verification
AU - Ladner, Tobias
AU - Althoff, Matthias
N1 - Publisher Copyright:
Copyright © 2024, Association for the Advancement of Artificial Intelligence (www.aaai.org). All rights reserved.
PY - 2024/3/25
Y1 - 2024/3/25
N2 - Formal verification of neural networks is a challenging problem due to the complexity and nonlinearity of neural networks. It has been shown that polynomial zonotopes can tightly enclose the output set of a neural network. Unfortunately, the tight enclosure comes with additional complexity in the set representation, thus, rendering subsequent operations expensive to compute, such as computing interval bounds and intersection checking. To address this issue, we present a novel approach to restructure a polynomial zonotope to tightly enclose the original polynomial zonotope while drastically reducing its complexity. The restructuring is achieved by relaxing the exponents of the dependent factors of polynomial zonotopes and finding an appropriate approximation error. We demonstrate the applicability of our approach on output sets of neural networks, where we obtain tighter results in various subsequent operations, such as order reduction, zonotope enclosure, and range bounding.
AB - Formal verification of neural networks is a challenging problem due to the complexity and nonlinearity of neural networks. It has been shown that polynomial zonotopes can tightly enclose the output set of a neural network. Unfortunately, the tight enclosure comes with additional complexity in the set representation, thus, rendering subsequent operations expensive to compute, such as computing interval bounds and intersection checking. To address this issue, we present a novel approach to restructure a polynomial zonotope to tightly enclose the original polynomial zonotope while drastically reducing its complexity. The restructuring is achieved by relaxing the exponents of the dependent factors of polynomial zonotopes and finding an appropriate approximation error. We demonstrate the applicability of our approach on output sets of neural networks, where we obtain tighter results in various subsequent operations, such as order reduction, zonotope enclosure, and range bounding.
UR - http://www.scopus.com/inward/record.url?scp=85189643898&partnerID=8YFLogxK
U2 - 10.1609/aaai.v38i19.30125
DO - 10.1609/aaai.v38i19.30125
M3 - Conference article
AN - SCOPUS:85189643898
SN - 2159-5399
VL - 38
SP - 21304
EP - 21311
JO - Proceedings of the AAAI Conference on Artificial Intelligence
JF - Proceedings of the AAAI Conference on Artificial Intelligence
IS - 19
T2 - 38th AAAI Conference on Artificial Intelligence, AAAI 2024
Y2 - 20 February 2024 through 27 February 2024
ER -