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

Most existing pre-trained language models for source code focus on learningthe static code text, typically augmented with static code structures (abstractsyntax tree, dependency graphs, etc.). However, program semantics will not befully exposed before the real execution. Without an understanding of theprogram execution, statically pre-trained models fail to comprehensivelycapture the dynamic code properties, such as the branch coverage and theruntime variable values, and they are consequently less effective at codeunderstanding tasks, such as retrieving semantic clones and detecting softwarevulnerabilities.  To close the gap between the static nature of language models and the dynamiccharacteristics of programs, we introduce TRACED, an execution-awarepre-training strategy for source code. Specifically, we pre-train code languagemodels with a combination of source code, executable inputs, and correspondingexecution traces. Our goal is to teach code models the complicated executionlogic during the pre-training, enabling the model to statically estimate thedynamic code properties without repeatedly executing code during task-specificfine-tuning.  To illustrate the effectiveness of our proposed approach, we fine-tune andevaluate TRACED on three downstream tasks: static execution estimation, cloneretrieval, and vulnerability detection. The empirical results show that TRACEDrelatively improves the statically pre-trained code models by 12.4% forcomplete execution path prediction and by 25.2% for runtime variable valuepredictions. TRACED also significantly outperforms statically pre-trainedmodels in clone retrieval and vulnerability detection across four publicbenchmarks.

TRACED: Execution-aware Pre-training for Source Code