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

Pretrained Language Models (PLMs) have emerged as the state-of-the-artparadigm for code search tasks. The paradigm involves pretraining the model onsearch-irrelevant tasks such as masked language modeling, followed by thefinetuning stage, which focuses on the search-relevant task. The typicalfinetuning method is to employ a dual-encoder architecture to encode semanticembeddings of query and code separately, and then calculate their similaritybased on the embeddings.  However, the typical dual-encoder architecture falls short in modelingtoken-level interactions between query and code, which limits the model'scapabilities. In this paper, we propose a novel approach to address thislimitation, introducing a cross-encoder architecture for code search thatjointly encodes the semantic matching of query and code. We further introduce aRetriever-Ranker (RR) framework that cascades the dual-encoder andcross-encoder to promote the efficiency of evaluation and online serving.Moreover, we present a probabilistic hard negative sampling method to improvethe cross-encoder's ability to distinguish hard negative codes, which furtherenhances the cascade RR framework. Experiments on four datasets using threecode PLMs demonstrate the superiority of our proposed method.

Retriever and Ranker Framework with Probabilistic Hard Negative Sampling for Code Search