Logical Segmentation of Source Code
This addresses a bottleneck in software analysis for developers and researchers by enabling more meaningful code decomposition beyond syntactic cues.
The paper tackles the problem of code segmentation by developing a novel deep learning approach that generates logical code segments independent of language or syntactic correctness, using a unique dataset construction technique to approximate ground truth. The method can improve tasks like automatic code commenting, vulnerability detection, bug repair, and code synthesis.
Many software analysis methods have come to rely on machine learning approaches. Code segmentation - the process of decomposing source code into meaningful blocks - can augment these methods by featurizing code, reducing noise, and limiting the problem space. Traditionally, code segmentation has been done using syntactic cues; current approaches do not intentionally capture logical content. We develop a novel deep learning approach to generate logical code segments regardless of the language or syntactic correctness of the code. Due to the lack of logically segmented source code, we introduce a unique data set construction technique to approximate ground truth for logically segmented code. Logical code segmentation can improve tasks such as automatically commenting code, detecting software vulnerabilities, repairing bugs, labeling code functionality, and synthesizing new code.