Word-Entity Duet Representations for Document Ranking

By Chenyan Xiong, Jamie Callan, Tie-Yan Liu. 2017

127 citations recorded by the source corpus.

This paper presents a word-entity duet framework for utilizing knowledge bases in ad-hoc retrieval. In this work, the query and documents are modeled by word-based representations and entity-based representations. Ranking features are generated by the interactions between the two representations, incorporating information from the word space, the entity space, and the cross-space connections through the knowledge graph. To handle the uncertainties from the automatically constructed entity representations, an attention-based ranking model AttR-Duet is developed. With back-propagation from ranking labels, the model learns simultaneously how to demote noisy entities and how to rank documents with the word-entity duet. Evaluation results on TREC Web Track ad-hoc task demonstrate that all of the four-way interactions in the duet are useful, the attention mechanism successfully steers the model away from noisy entities, and together they significantly outperform both word-based and entity-based learning to rank systems.

Subfields: Artificial Intelligence, Information Systems, Information Retrieval & Search

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Information Systems

In Information Retrieval & Search

Citations in the corpus

XLNet: Generalized Autoregressive Pretraining for Language Understanding (2019, background): “55 in ERR@20, slightly worse than ours) in [36] as it uses additional entity-related data.”
Multi-step Entity-centric Information Retrieval for Multi-Hop Question Answering (2019, neutral): “There has been a lot of work which leverages knowledge graphs (KGs) to learn better entity representations (Xiong and Callan, 2015; Xiong et al., 2017; Liu et al., 2018) and for better query reformulation (Cao et al.”
Knowledge-Aware Language Model Pretraining (2020, methodology): “The first step to introduce knowledge awareness is an entity tokenizer that forms an additional entity token sequence [26] to signal the existence of entities in the input and output of the pretraining process.”
REL: An Entity Linker Standing on the Shoulders of Giants (2020, background): “Over the past years, entity linking has become a standard component in modern retrieval systems, and has been leveraged in a range of tasks, including document ranking [26], entity retrieval [11], knowledge base population [3], and”
Ranking of text documents using TF-IDF weighting and association rules mining (2018, methodology): “This paper [12] presents a word-entity duet framework for utilizing knowledge bases in ad-hoc retrieval.”
A Survey on Application of Knowledge Graph (2020, background): “[35] contribute to capturing more semantic relevance patterns.”
Selective Weak Supervision for Neural Information Retrieval (2020, methodology): “All experiment settings are kept consistent with Dai and Callan [4] (Xiong et al. [44] on ClueWeb12).”
A Deep Look into Neural Ranking Models for Information Retrieval (2019, background): “external knowledge [29, 30], and other novel uses of neural approaches for IR tasks [31, 32].”
Convolutional Neural Networks for Soft-Matching N-Grams in Ad-hoc Search (2018, unknown): “We followed a standard re-ranking methodology in prior research [8, 28]: re-rank the top 100 candidate documents retrieved by Galago using sequential dependency model queries; the INQUERY stopword list augmented with web-specic stop words; KStemming; and spam ltering using Waterloo spam score with”
Efficient Interaction-based Neural Ranking with Locality Sensitive Hashing (2019, methodology): “Neural networks have been applied in information retrieval and text mining tasks (e.g. [5, 9, 20, 23, 24, 33–36]).”