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

Deploying dense retrieval models efficiently is becoming increasinglyimportant across various industries. This is especially true for enterprisesearch services, where customizing search engines to meet the time demands ofdifferent enterprises in different domains is crucial. Motivated by this, wedevelop a time-efficient approach called DREditor to edit the matching rule ofan off-the-shelf dense retrieval model to suit a specific domain. This isachieved by directly calibrating the output embeddings of the model using anefficient and effective linear mapping. This mapping is powered by an editoperator that is obtained by solving a specially constructed least squaresproblem. Compared to implicit rule modification via long-time finetuning, ourexperimental results show that DREditor provides significant advantages ondifferent domain-specific datasets, dataset sources, retrieval models, andcomputing devices. It consistently enhances time efficiency by 100-300 timeswhile maintaining comparable or even superior retrieval performance. In abroader context, we take the first step to introduce a novel embeddingcalibration approach for the retrieval task, filling the technical blank in thecurrent field of embedding calibration. This approach also paves the way forbuilding domain-specific dense retrieval models efficiently and inexpensively.

DREditor: An Time-efficient Approach for Building a Domain-specific Dense Retrieval Model