@inproceedings{8d97609c15b1464a92d2091aa3000823,
title = "POLAR: A Polynomial Arithmetic Framework for Verifying Neural-Network Controlled Systems",
abstract = "We present POLAR (The source code can be found at https://github.com/ChaoHuang2018/POLAR_Tool. The full version of this paper can be found at https://arxiv.org/abs/2106.13867. ), a POLynomial ARithmetic-based framework for efficient time-bounded reachability analysis of neural-network controlled systems. Existing approaches leveraging the standard Taylor Model (TM) arithmetic for approximating the neural-network controller cannot deal with non-differentiable activation functions and suffer from rapid explosion of the remainder when propagating TMs. POLAR overcomes these shortcomings by integrating TM arithmetic with Bernstein polynomial interpolation and symbolic remainders. The former enables TM propagation across non-differentiable activation functions and local refinement of TMs, and the latter reduces error accumulation in the TM remainder for linear mappings in the neural network. Experimental results show POLAR significantly outperforms the state-of-the-art tools on both efficiency and tightness of the reachable set overapproximation.",
author = "Chao Huang and Jiameng Fan and Xin Chen and Wenchao Li and Qi Zhu",
note = "Publisher Copyright: {\textcopyright} 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.; 20th International Symposium on Automated Technology for Verification and Analysis, ATVA 2022 ; Conference date: 25-10-2022 Through 28-10-2022",
year = "2022",
doi = "10.1007/978-3-031-19992-9_27",
language = "English (US)",
isbn = "9783031199912",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Science and Business Media Deutschland GmbH",
pages = "414--430",
editor = "Ahmed Bouajjani and Luk{\'a}{\v s} Hol{\'i}k and Zhilin Wu",
booktitle = "Automated Technology for Verification and Analysis - 20th International Symposium, ATVA 2022, Proceedings",
address = "Germany",
}