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

Passage retrieval aims to retrieve relevant passages from large collectionsof the open-domain corpus. Contextual Masked Auto-Encoding has been proveneffective in representation bottleneck pre-training of a monolithicdual-encoder for passage retrieval. Siamese or fully separated dual-encodersare often adopted as basic retrieval architecture in the pre-training andfine-tuning stages for encoding queries and passages into their latentembedding spaces. However, simply sharing or separating the parameters of thedual-encoder results in an imbalanced discrimination of the embedding spaces.In this work, we propose to pre-train Contextual Masked Auto-Encoder withMixture-of-Textual-Experts (CoT-MoTE). Specifically, we incorporatetextual-specific experts for individually encoding the distinct properties ofqueries and passages. Meanwhile, a shared self-attention layer is still keptfor unified attention modeling. Results on large-scale passage retrievalbenchmarks show steady improvement in retrieval performances. The quantitiveanalysis also shows a more balanced discrimination of the latent embeddingspaces.

CoT-MoTE: Exploring ConTextual Masked Auto-Encoder Pre-training with Mixture-of-Textual-Experts for Passage Retrieval