Multimodal Representation for Neural Code Search

TL;DR

This paper introduces a multimodal representation approach using tree-serialization of ASTs to enhance semantic code search, demonstrating improved performance across a large, multi-language dataset.

Contribution

It proposes a novel multimodal code representation method with tree-serialization of ASTs, advancing semantic code search accuracy.

Findings

Tree-serialized representations improve code search performance.

Multimodal learning models outperform existing methods.

Metrics for semantic and syntactic completeness aid understanding.

Abstract

Semantic code search is about finding semantically relevant code snippets for a given natural language query. In the state-of-the-art approaches, the semantic similarity between code and query is quantified as the distance of their representation in the shared vector space. In this paper, to improve the vector space, we introduce tree-serialization methods on a simplified form of AST and build the multimodal representation for the code data. We conduct extensive experiments using a single corpus that is large-scale and multi-language: CodeSearchNet. Our results show that both our tree-serialized representations and multimodal learning model improve the performance of code search. Last, we define intuitive quantification metrics oriented to the completeness of semantic and syntactic information of the code data, to help understand the experimental findings.