A Method of Vulnerability Assessment Based on Deep Learning and OSS Fault Big Data

Software vulnerability is generally defined as the weakness of security caused by the fault. Waterfall model has been usually used for the software development till recent past. Also, a number of open source components have been implemented in many commercial software. Recently, the open source software have extended to the cloud service and edge computing and the big data. It is thus imperetive to consider the impacts from the big data and network access. Various vulnerability assessment methods have been proposed by several researchers. In the typical vulnerability problems, methods of vulnerability assessment considering the fault factors have not been presented. Although, it is difficult to assess many fault factors recorded on the bug tracking system because of the uncertainty. The authors, in this paper, propose an assessment method of vulnerability by using the deep learning. Moreover, actual data to showcase the numerical examples for the estimation method of unknown parameters included in the proposed model for the vulnerability assessment have been presented.