Source Code Authorship Attribution Using Long Short-Term Memory Based Networks

Abstract

Machine learning approaches to source code authorship attribution attempt to find statistical regularities in human-generated source code that can identify the author or authors of that code. This has applications in plagiarism detection, intellectual property infringement, and post-incident forensics in computer security. The introduction of features derived from the Abstract Syntax Tree (AST) of source code has recently set new benchmarks in this area, significantly improving over previous work that relied on easily obfuscatable lexical and format features of program source code. However, these AST-based approaches rely on hand-constructed features derived from such trees, and often include ancillary information such as function and variable names that may be obfuscated or manipulated. In this work, we provide novel contributions to AST-based source code authorship attribution using deep neural networks. We implement Long Short-Term Memory (LSTM) and Bidirectional Long Short-Term Memory (BiLSTM) models to automatically extract relevant features from the AST representation of programmers’ source code. We show that our models can automatically learn efficient representations of AST-based features without needing hand-constructed ancillary information used by previous methods. Our empirical study on multiple datasets with different programming languages shows that our proposed approach achieves the state-of-the-art performance for source code authorship attribution on AST-based features, despite not leveraging information that was previously thought to be required for high-confidence classification.