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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Building deep learning models on source code has found many successfulsoftware engineering applications, such as code search, code commentgeneration, bug detection, code migration, and so on. Current learningtechniques, however, have a major drawback that these models are mostly trainedon datasets labeled for particular downstream tasks, and code representationsmay not be suitable for other tasks. While some techniques producerepresentations from unlabeled code, they are far from satisfactory whenapplied to downstream tasks. Although certain techniques generaterepresentations from unlabeled code when applied to downstream tasks they arefar from satisfactory. This paper proposes InferCode to overcome the limitationby adapting the self-supervised learning mechanism to build source code model.The key novelty lies in training code representations by predictingautomatically identified subtrees from the context of the ASTs. Subtrees inASTs are treated with InferCode as the labels for training code representationswithout any human labeling effort or the overhead of expensive graphconstruction, and the trained representations are no longer tied to anyspecific downstream tasks or code units. We trained an InferCode model instanceusing the Tree-based CNN as the encoder of a large set of Java code and appliedit to downstream unsupervised tasks such as code clustering, code clonedetection, cross-language code search or reused under a transfer learningscheme to continue training the model weights for supervised tasks such as codeclassification and method name prediction. Compared to previous code learningtechniques applied to the same downstream tasks, such as Code2Vec, Code2Seq,ASTNN, higher performance results are achieved using our pre-trained InferCodemodel with a significant margin for most tasks including those involvingdifferent programming languages.

InferCode: Self-Supervised Learning of Code Representations by Predicting Subtrees