English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Much of software-engineering research relies on the naturalness of code, thefact that code, in small code snippets, is repetitive and can be predictedusing statistical language models like n-gram. Although powerful, training suchmodels on large code corpus is tedious, time-consuming and sensitive to codepatterns (and practices) encountered during training. Consequently, thesemodels are often trained on a small corpora and estimate the languagenaturalness that is relative to a specific style of programming or type ofproject. To overcome these issues, we propose using pre-trained language modelsto infer code naturalness. Pre-trained models are often built on big data, areeasy to use in an out-of-the-box way and include powerful learning associationsmechanisms. Our key idea is to quantify code naturalness through itspredictability, by using state-of-the-art generative pre-trained languagemodels. To this end, we infer naturalness by masking (omitting) code tokens,one at a time, of code-sequences, and checking the models' ability to predictthem. To this end, we evaluate three different predictability metrics; a)measuring the number of exact matches of the predictions, b) computing theembedding similarity between the original and predicted code, i.e., similarityat the vector space, and c) computing the confidence of the model when doingthe token completion task irrespective of the outcome. We implement thisworkflow, named CodeBERT-nt, and evaluate its capability to prioritize buggylines over non-buggy ones when ranking code based on its naturalness. Ourresults, on 2510 buggy versions of 40 projects from the SmartShark dataset,show that CodeBERT-nt outperforms both, random-uniform and complexity-basedranking techniques, and yields comparable results (slightly better) than then-gram models.

CodeBERT-nt: code naturalness via CodeBERT