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 Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Knowledge-intensive language tasks (KILT) usually require a large body ofinformation to provide correct answers. A popular paradigm to solve thisproblem is to combine a search system with a machine reader, where the formerretrieves supporting evidences and the latter examines them to produce answers.Recently, the reader component has witnessed significant advances with the helpof large-scale pre-trained generative models. Meanwhile most existing solutionsin the search component rely on the traditional ``index-retrieve-then-rank''pipeline, which suffers from large memory footprint and difficulty inend-to-end optimization. Inspired by recent efforts in constructing model-basedIR models, we propose to replace the traditional multi-step search pipelinewith a novel single-step generative model, which can dramatically simplify thesearch process and be optimized in an end-to-end manner. We show that a stronggenerative retrieval model can be learned with a set of adequately designedpre-training tasks, and be adopted to improve a variety of downstream KILTtasks with further fine-tuning. We name the pre-trained generative retrievalmodel as CorpusBrain as all information about the corpus is encoded in itsparameters without the need of constructing additional index. Empirical resultsshow that CorpusBrain can significantly outperform strong baselines for theretrieval task on the KILT benchmark and establish new state-of-the-artdownstream performances. We also show that CorpusBrain works well under zero-and low-resource settings.

CorpusBrain: Pre-train a Generative Retrieval Model for Knowledge-Intensive Language Tasks