Optimizing Secure Software Development Lifecycle with Parallel Temporal Convolutional Network for Vulnerability Detection

Smart contracts are one of the software application which is developed based on blockchain, are extensively used in Internet of Things, finance management, and other applications. However, development of smart contracts increases, meanwhile vulnerability attacks elated to the software also increases which occur often. To overcome this limitation, Parallel Temporal Convolutional Network (PTCN) is proposed for the vulnerability detection in secure software development. By using parallel temporal convolutions, the detection model process multiple time steps at the same time, that helps to identify patterns effectively. Firstly, Smartbugs datasets are used for acquiring data regarding the vulnerable behaviors in a software development. After that these data are pre-processed by regularization method and fed to word embedding for converting text data to vector data. Finally, from these vector data, the proposed PTCN model detects the vulnerability efficiently. The experimental results of the proposed PTCN method attained accuracy of 93.26% which is higher than the existing detection methods such as Support Vector Machine (SVM) and Convolutional Neural Network (CNN).