Synthesizing ReLU neural networks with two hidden layers as barrier certificates for hybrid systems
Authors:
Open Access
- 19 May 2021
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 1-11
Abstract
Barrier certificates provide safety guarantees for hybrid systems. In this paper, we propose a novel approach to synthesizing neural networks as barrier certificates. Candidate networks are trained from a special structure: ReLU neural networks consisting of two hidden layers. Then, the problem of identifying real barrier certificates from candidates is transformed into a group of mixed integer linear programming problems and a mixed integer quadratically constrained problem. Taking full advantage of the recent advance in optimization, barrier certificates validation can be performed effectively. We implement the tool SyntheBC and evaluate its performance over 3 hybrid systems and 8 continuous systems up to 12-dimensional state space. The experimental results show that our method is more scalable and effective than the classical polynomial barrier certificate method and the existing neural network based method.Keywords
Funding Information
- National Natural Science Foundation of China (62032010, 61772203, 61772260)
- Zhejiang Provincial Natural Science Foundation of China (LY20F020020)
- Jiangsu Natural Science Foundation (BK20202001)
This publication has 27 references indexed in Scilit:
- A New Barrier Certificate for Safety Verification of Hybrid SystemsThe Computer Journal, 2013
- Exponential-Condition-Based Barrier Certificate Generation for Safety Verification of Hybrid Systems2013
- Computing differential invariants of hybrid systems as fixedpointsFormal Methods in System Design, 2009
- Constraint-Based Approach for Analysis of Hybrid Systems2008
- Constructing invariants for hybrid systemsFormal Methods in System Design, 2007
- Generating Polynomial Invariants for Hybrid SystemsLecture Notes in Computer Science, 2005
- Safety Verification of Hybrid Systems Using Barrier Certificates2004
- Reach Set Computations Using Real Quantifier Elimination2001
- Approximation by superpositions of a sigmoidal functionMathematics of Control, Signals, and Systems, 1989
- Multilayer feedforward networks are universal approximatorsNeural Networks, 1989