Automated and Formal Synthesis of Neural Barrier Certificates for Dynamical Models

被引:25
|
作者
Peruffo, Andrea [1 ]
Ahmed, Daniele [2 ]
Abate, Alessandro [1 ]
机构
[1] Univ Oxford, Dept Comp Sci, Oxford, England
[2] Amazon Inc, London, England
来源
TOOLS AND ALGORITHMS FOR THE CONSTRUCTION AND ANALYSIS OF SYSTEMS, PT I, TACAS 2021 | 2021年 / 12651卷
关键词
HYBRID SYSTEMS; SAFETY VERIFICATION;
D O I
10.1007/978-3-030-72016-2_20
中图分类号
TP31 [计算机软件];
学科分类号
081202 ; 0835 ;
摘要
We introduce an automated, formal, counterexample-based approach to synthesise Barrier Certificates (BC) for the safety verification of continuous and hybrid dynamical models. The approach is underpinned by an inductive framework: this is structured as a sequential loop between a learner, which manipulates a candidate BC structured as a neural network, and a sound verifier, which either certifies the candidate's validity or generates counter-examples to further guide the learner. We compare the approach against state-of-the-art techniques, over polynomial and non-polynomial dynamical models: the outcomes show that we can synthesise sound BCs up to two orders of magnitude faster, with in particular a stark speedup on the verification engine (up to three orders less), whilst needing a far smaller data set (up to three orders less) for the learning part. Beyond improvements over the state of the art, we further challenge the new approach on a hybrid dynamical model and on larger-dimensional models, and showcase the numerical robustness of our algorithms and codebase.
引用
收藏
页码:370 / 388
页数:19
相关论文
共 50 条
  • [31] Scalable Synthesis of Safety Barrier Certificates for Networks of Stochastic Switched Systems
    Lavaei, Abolfazl
    Frazzoli, Emilio
    IEEE TRANSACTIONS ON AUTOMATIC CONTROL, 2024, 69 (11) : 7294 - 7309
  • [32] The application of automated reasoning to formal models of combinatorial optimization
    Helman, P
    Veroff, R
    APPLIED MATHEMATICS AND COMPUTATION, 2001, 120 (1-3) : 175 - 194
  • [33] An automated tool for semantic accessing to formal software models
    Wang, Hai H.
    Damljanovic, Danica
    Sun, Jing
    SCIENCE OF COMPUTER PROGRAMMING, 2014, 95 : 93 - 111
  • [34] Formal languages and neural models for learning on sequences
    Merrill, William
    INTERNATIONAL CONFERENCE ON GRAMMATICAL INFERENCE, VOL 217, 2023, 217 : 5 - 5
  • [35] Formal derivation of qualitative dynamical models from biochemical networks
    Abou-Jaoude, Wassim
    Thieffry, Denis
    Feret, Jerome
    BIOSYSTEMS, 2016, 149 : 70 - 112
  • [36] Automated synthesis of prediction models for neural network based myocardial infarction classifiers
    Lopez, JA
    Nugent, CD
    Black, ND
    Smith, AE
    PROCEEDINGS OF THE 23RD ANNUAL INTERNATIONAL CONFERENCE OF THE IEEE ENGINEERING IN MEDICINE AND BIOLOGY SOCIETY, VOLS 1-4: BUILDING NEW BRIDGES AT THE FRONTIERS OF ENGINEERING AND MEDICINE, 2001, 23 : 3803 - 3806
  • [37] Synthesizing ReLU Neural Networks with Two Hidden Layers as Barrier Certificates for Hybrid Systems
    Zhao, Qingye
    Chen, Xin
    Zhang, Yifan
    Sha, Meng
    Yang, Zhengfeng
    Lin, Wang
    Tang, Enyi
    Chen, Qiguang
    Li, Xuandong
    HSCC2021: PROCEEDINGS OF THE 24TH INTERNATIONAL CONFERENCE ON HYBRID SYSTEMS: COMPUTATION AND CONTROL (PART OF CPS-IOT WEEK), 2021,
  • [38] Active Sampling based Safe Identification of Dynamical Systems using Extreme Learning Machines and Barrier Certificates
    Salehi, Iman
    Yao, Gang
    Dani, Ashwin P.
    2019 INTERNATIONAL CONFERENCE ON ROBOTICS AND AUTOMATION (ICRA), 2019, : 22 - 28
  • [39] Constrained Block Nonlinear Neural Dynamical Models
    Skomski, Elliott
    Vasisht, Soumya
    Wight, Colby
    Tuor, Aaron
    Drgona, Jan
    Vrabie, Draguna
    2021 AMERICAN CONTROL CONFERENCE (ACC), 2021, : 3993 - 4000
  • [40] Data-driven verification and synthesis of stochastic systems via barrier certificates
    Salamati, Ali
    Lavaei, Abolfazl
    Soudjani, Sadegh
    Zamani, Majid
    AUTOMATICA, 2024, 159