Pre-trained Language Model based Ranking in Baidu Search

Zou, Lixin; Zhang, Shengqiang; Cai, Hengyi; Ma, Dehong; Cheng, Suqi; Shi, Daiting; Zhu, Zhifan; Su, Weiyue; Wang, Shuaiqiang; Cheng, Zhicong; Yin, Dawei

doi:10.48550/arxiv.2105.11108

Cited by 1 publication

(1 citation statement)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This role is usually referred to as first-stage retrieval or candidate generation. Multi-stage ranking architectures have been studied extensively by academic researchers [Matveeva et al, 2006, Cambazoglu et al, 2010, Wang et al, 2011, Tonellotto et al, 2013, Asadi and Lin, 2013, Capannini et al, 2016, Clarke et al, 2016, Chen et al, 2017, Mackenzie et al, 2018 and there is substantial documentation that many commercial applications are designed in this manner [Pedersen, 2010, Liu et al, 2017, Huang et al, 2020, Zou et al, 2021.…”

Section: Introductionmentioning

confidence: 99%

A Proposed Conceptual Framework for a Representational Approach to Information Retrieval

Lin¹

2021

Preprint

View full text Add to dashboard Cite

This paper outlines a conceptual framework for understanding recent developments in information retrieval and natural language processing that attempts to integrate dense and sparse retrieval methods. I propose a representational approach that breaks the core text retrieval problem into a logical scoring model and a physical retrieval model. The scoring model is defined in terms of encoders, which map queries and documents into a representational space, and a comparison function that computes query-document scores. The physical retrieval model defines how a system produces the top-k scoring documents from an arbitrarily large corpus with respect to a query. The scoring model can be further analyzed along two dimensions: dense vs. sparse representations and supervised (learned) vs. unsupervised approaches. I show that many recently proposed retrieval methods, including multi-stage ranking designs, can be seen as different parameterizations in this framework, and that a unified view suggests a number of open research questions, providing a roadmap for future work. As a bonus, this conceptual framework establishes connections to sentence similarity tasks in natural language processing and information access "technologies" prior to the dawn of computing.

show abstract

Section: Introductionmentioning

confidence: 99%

A Proposed Conceptual Framework for a Representational Approach to Information Retrieval

Lin¹

2021

Preprint

View full text Add to dashboard Cite

show abstract

Pre-trained Language Model based Ranking in Baidu Search

Cited by 1 publication

References 36 publications

A Proposed Conceptual Framework for a Representational Approach to Information Retrieval

A Proposed Conceptual Framework for a Representational Approach to Information Retrieval

Contact Info

Product

Resources

About