作者: Zhaoyan Jin , Yang Yu
关键词:
摘要: The quantity and volume of software increase rapidly, more vulnerabilities are hidden there, thus vulnerability detection becomes important. This paper reviews works about machine learning based methods for source code, especially on program representation vectorization, Machine Learning Methods. However, current detect in coarse-grained level, locating the still needs much additional work. problem can be solved by fine-grained level. assumes that level will a future research trend, proposes several possible solutions.