版权所有:内蒙古大学图书馆 技术提供:维普资讯• 智图
内蒙古自治区呼和浩特市赛罕区大学西街235号 邮编: 010021
作者机构:Dalian Univ Technol Sch Software Technol Dalian 116621 Peoples R China Key Lab Ubiquitous Network & Serv Software Liaonin Dalian 116621 Peoples R China
出 版 物:《INTERNATIONAL JOURNAL OF SOFTWARE ENGINEERING AND KNOWLEDGE ENGINEERING》 (Int. J. Software Engineer. Knowledge Engineer.)
年 卷 期:2024年第34卷第8期
页 面:1199-1220页
核心收录:
学科分类:0808[工学-电气工程] 08[工学] 0835[工学-软件工程] 0812[工学-计算机科学与技术(可授工学、理学学位)]
主 题:Neural networks robustness verification formal methods dual-neuron relaxation
摘 要:In the robustness verification of neural networks, formal methods have been used to give deterministic guarantees for neural networks. However, recent studies have found that the verification method of single-neuron relaxation in this field has an inherent convex barrier that affects its verification capability. To address this problem, we propose a new verification method by combining dual-neuron relaxation and linear programming. This method captures the dependencies between different neurons in the same hidden layer by adding a two-neuron joint constraint to the linear programming model, thus overcoming the convex barrier problem caused by relaxation for only a single neuron. Our method avoids the combination of exponential inequality constraints and can be computed in polynomial time. Experimental results show that we can obtain tighter bounds and achieve more accurate verification than single-neuron relaxation methods.