KFCC: A differentiation-aware and keyword-guided fine-grain code comment generation model

• DOI: 10.1016/j.eswa.2024.123946
• Publication Date: 2024-04-11T01:07:08Z
• Authors: Rui Zhang; Ziyue Qiao; Chenghao Zhang; Jianjun Yu
• Keywords: Code (set theory)

An efficient and accurate understanding of the intent code is an indispensable skill in computer technology, especially collaborative engineering experimental reproduction. AI-assisted automated comment generator, with goal generating programmer-readable explanations, has been emerging hot topic for software project comprehension. Despite promising performances, three critical issues emerged: 1) The summary limited fine-grain details code. 2) key-word level guidance model should be included better comments generation. 3) performance generative may dampened by noises manual annotation. In response, we propose a novel generation, scenario statement-level assistance method-level comment. We also KFCC, differentiation-aware keyword-guided generation model. Specifically, proposed KFCC generates incorporating key information extracted keyword extractor gate fusion way. To enhance effectiveness robustness model, enhancing encoder comprehension, letting distinguish significant knowledge via contrastive learning. Extensive experiments conducted on open-source projects demonstrate that achieves outstanding six programming languages (including Ruby, Python, JavaScript, Java, etc.) CodeSearchNet benchmark.