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 Tools annual

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 monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Deep prompt tuning (DPT) has gained great success in most natural languageprocessing~(NLP) tasks. However, it is not well-investigated in dense retrievalwhere fine-tuning~(FT) still dominates. When deploying multiple retrieval tasksusing the same backbone model~(e.g., RoBERTa), FT-based methods are unfriendlyin terms of deployment cost: each new retrieval model needs to repeatedlydeploy the backbone model without reuse. To reduce the deployment cost in sucha scenario, this work investigates applying DPT in dense retrieval. Thechallenge is that directly applying DPT in dense retrieval largelyunderperforms FT methods. To compensate for the performance drop, we proposetwo model-agnostic and task-agnostic strategies for DPT-based retrievers,namely retrieval-oriented intermediate pretraining and unified negative mining,as a general approach that could be compatible with any pre-trained languagemodel and retrieval task. The experimental results show that the proposedmethod (called DPTDR) outperforms previous state-of-the-art models on bothMS-MARCO and Natural Questions. We also conduct ablation studies to examine theeffectiveness of each strategy in DPTDR. We believe this work facilitates theindustry, as it saves enormous efforts and costs of deployment and increasesthe utility of computing resources. Our code is available athttps://github.com/tangzhy/DPTDR.

DPTDR: Deep Prompt Tuning for Dense Passage Retrieval