Software Security Testing through Coverage in Deep Neural Networks

Abstract

With the continuous progress of society, computer technology and information technology are also experiencing rapid development. Especially in recent years, the application of computer technology has rapidly entered into people's daily life. As people’s lives become richer, these applications have become particularly complex. For some large software, tens of thousands of function points or millions of lines of source code may be triggered to support it when performing related tasks. As a result, the security of such a complicated and excellent software becomes quite essential. The most effective way to ensure software security is to test the security of software products during the development process. A precise and effective security testing process is the basis for ensuring that software is tested for security. Without a detailed scientific software security testing model to guide software development for security testing, software security testing will become very difficult. This not only wastes more time and money but also does not guarantee the security of the software. A great security testing methodology should be able to find security problems that may be hidden deep within the software. In addition, a scientific process management can greatly facilitate the implementation of software security testing. As a result, it is relatively meaningful to establish a complete software security testing process model, generate excellent security test cases, and develop security process management tools for software security testing. At the same time, in recent years, deep learning has gradually entered more and more people's lives. However, the widespread application of deep learning systems can bring convenience to human life but also bring some hidden dangers. Hence, deep neural networks must be adequately tested to eliminate as many security risks as possible in some safety-critical software that involves personal and property safety. As the foundation of deep learning systems, deep neural networks should be adequately tested for security. However, deep learning systems are fundamentally different from traditional software testing, so traditional software testing techniques cannot be directly applied to deep neural network testing. In recent years, many scholars in related fields have proposed coverage guidelines based on deep learning testing, but the usefulness of these guidelines is still debatable. Based on the complexity of the large software development process and the fact that the interrelationship between nodes often constitutes a complex network of collaborative relationships, this study applies coverage-based testing in deep neural networks to test the security of software. To be specific, this research applies metrics such as peak coverage, speed to peak, and computational speed to evaluate coverage criteria and to investigate the feasibility of using coverage to guide test case selection to select solutions for security testing.