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

Numerous code changes are made by developers in their daily work, and asuperior representation of code changes is desired for effective code changeanalysis. Recently, Hoang et al. proposed CC2Vec, a neural network-basedapproach that learns a distributed representation of code changes to capturethe semantic intent of the changes. Despite demonstrated effectiveness inmultiple tasks, CC2Vec has several limitations: 1) it considers onlycoarse-grained information about code changes, and 2) it relies on log messagesrather than the self-contained content of the code changes. In this work, wepropose CCBERT (\underline{C}ode \underline{C}hange \underline{BERT}), a newTransformer-based pre-trained model that learns a generic representation ofcode changes based on a large-scale dataset containing massive unlabeled codechanges. CCBERT is pre-trained on four proposed self-supervised objectives thatare specialized for learning code change representations based on the contentsof code changes. CCBERT perceives fine-grained code changes at the token levelby learning from the old and new versions of the content, along with the editactions. Our experiments demonstrate that CCBERT significantly outperformsCC2Vec or the state-of-the-art approaches of the downstream tasks by7.7\%--14.0\% in terms of different metrics and tasks. CCBERT consistentlyoutperforms large pre-trained code models, such as CodeBERT, while requiring6--10$\times$ less training time, 5--30$\times$ less inference time, and7.9$\times$ less GPU memory.

CCBERT: Self-Supervised Code Change Representation Learning