PACRR: A Position-Aware Neural IR Model for Relevance Matching

By Kai Hui, Andrew Yates, Klaus Berberich, Gerard de Melo. 2017

165 citations recorded by the source corpus.

In order to adopt deep learning for information retrieval, models are needed that can capture all relevant information required to assess the relevance of a document to a given user query. While previous works have successfully captured unigram term matches, how to fully employ position-dependent information such as proximity and term dependencies has been insufficiently explored. In this work, we propose a novel neural IR model named PACRR aiming at better modeling position-dependent interactions between a query and a document. Extensive experiments on six years’ TREC Web Track data confirm that the proposed model yields better results under multiple benchmarks.

Subfields: Artificial Intelligence, Information Systems, Signal Processing

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 Signal Processing

Citations in the corpus

Pretrained Transformers for Text Ranking: BERT and Beyond (2020, neutral): “, 2017], and PACRR [Hui et al., 2017].”
Efficient Document Re-Ranking for Transformers by Precomputing Term Representations (2020, methodology): “he efficiency of such networks. 2.1 Neural Ranking As neural approaches have gained prominence in other disciplines, many have investigated how deep neural networks can be applied to document ranking [10, 17, 19, 44]. These approaches typically act as a final-stage ranking function, via atelescoping”
Interpretable & Time-Budget-Constrained Contextualization for Re-Ranking (2020, methodology): “rary scoring results more feasible compared to pattern based scoring methods (for example PACRR [15]).”
Capreolus: A Toolkit for End-to-End Neural Ad Hoc Retrieval (2020, methodology): “• PACRR.”
TREC Complex Answer Retrieval Overview (2018, background): “• The MPIID5 submission includes three runs for passage retrieval, all of which are extensions of the PACRR [5] deep neural relevance ranking architecture: nn6pos, nn4poshperc, and nn6postprob.”
AUEB at BioASQ 7: Document and Snippet Retrieval (2019, methodology): “[15] explored ways to combine elmo [22] and bert [5] with complex neural ir models such as drmm [7] and pacrr [9].”
Retrieve-and-Read: Multi-task Learning of Information Retrieval and Reading Comprehension (2018, background): “There are roughly two groups: embedding space models [19, 32, 41] and interaction-based models [10, 16, 20, 21, 51].”
Zero-shot Neural Retrieval via Domain-targeted Synthetic Query Generation (2020, background): “As a result, neural retrieval models achieve new state-of-the-art performance on several information retrieval (IR) benchmarks (Guo et al., 2016; Pang et al., 2016; Hui et al., 2017; Dai et al., 2018; MacAvaney et al., 2019; Yang et al., 2019a; Gillick et al., 2018; Yang et al., 2019b).”
Teaching a New Dog Old Tricks: Resurrecting Multilingual Retrieval Using Zero-Shot Learning (2019, methodology): “More recently, a new wave of supervised state-of-the-art ranking models have been proposed by researchers [11,14,21,24,26,35,37]; these models rely on neural architectures to rerank the head of search results retrieved using a traditional unsupervised ranking algorithm, such as BM25.”
Leveraging Passage-level Cumulative Gain for Document Ranking (2020, background): “Position-Aware Convolutional Recurrent Relevance Matching (PACRR) [13] uses convolutional layers to capture both term matching and positional information based on the query-document interactions.”
PROP: Pre-training with Representative Words Prediction for Ad-hoc Retrieval (2020, methodology): “existing neural ranking models such as DRMM [14], PACRR [17] and KNRM [39].”
Let's measure run time! Extending the IR replicability infrastructure to include performance aspects (2019, methodology): “MatchPyramid and PACRR exhibit similar performance.”
An anatomy for neural search engines (2019, methodology): “Therefore, the authors opted for implementing what they judged to be the most consolidated, accessible and easily reproducible neural models in the literature, which are also a reference for several recent works [McDonald et al., 2018; Hui et al., 2017; Pang et al., 2017b].”
Promptagator: Few-shot Dense Retrieval From 8 Examples (2022, methodology): “Interaction based models on the other hand explicitly model the interaction between query and document terms (Guo et al., 2016; Hui et al., 2017; Xiong et al., 2017; Dai et al., 2018; McDonald et al., 2018; Nogueira & Cho, 2019), and therefore make more informed decisions.”