DeepRank: A New Deep Architecture for Relevance Ranking in Information Retrieval

Abstract

This paper concerns a deep learning approach to relevance ranking in information retrieval (IR). Existing deep IR models such as DSSM and CDSSM directly apply neural networks to generate ranking scores, without explicit understandings of the relevance. According to the human judgement process, a relevance label is generated by the following three steps: 1) relevant locations are detected, 2) local relevances are determined, 3) local relevances are aggregated to output the relevance label. In this paper we propose a new deep learning architecture, namely DeepRank, to simulate the above human judgment process. Firstly, a detection strategy is designed to extract the relevant contexts. Then, a measure network is applied to determine the local relevances by utilizing a convolutional neural network (CNN) or two-dimensional gated recurrent units (2D-GRU). Finally, an aggregation network with sequential integration and term gating mechanism is used to produce a global relevance score. DeepRank well captures important IR characteristics, including exact/semantic matching signals, proximity heuristics, query term importance, and diverse relevance requirement. Experiments on both benchmark LETOR dataset and a large scale clickthrough data show that DeepRank can significantly outperform learning to ranking methods, and existing deep learning methods.

Published as a conference paper at CIKM 2017, CIKM'17, November 6--10, 2017, Singapore TextNet (https://github.com/pl8787/textnet-release) PyTorch (https://github.com/pl8787/DeepRank_PyTorch)

References in corpus (1)

• A Deep Relevance Matching Model for Ad-hoc Retrieval

Cited by in corpus (27)

• Deeper Text Understanding for IR with Contextual Neural Language Modeling
• Information Retrieval: Recent Advances and Beyond
• Modeling Diverse Relevance Patterns in Ad-hoc Retrieval
• Representation Learning for Natural Language Processing
• Learning Groupwise Multivariate Scoring Functions Using Deep Neural Networks
• PROP: Pre-training with Representative Words Prediction for Ad-hoc Retrieval
• Neural ranking models for document retrieval
• Overview of BioASQ 2020: The eighth BioASQ challenge on Large-Scale Biomedical Semantic Indexing and Question Answering
• Generic Intent Representation in Web Search
• StruBERT: Structure-aware BERT for Table Search and Matching
• Learning a Product Relevance Model from Click-Through Data in E-Commerce
• Feature-aware Diversified Re-ranking with Disentangled Representations for Relevant Recommendation
• Selective Weak Supervision for Neural Information Retrieval
• Match-Ignition: Plugging PageRank into Transformer for Long-form Text Matching
• Listwise Learning to Rank by Exploring Unique Ratings
• Match-Prompt: Improving Multi-task Generalization Ability for Neural Text Matching via Prompt Learning
• The Power of Selecting Key Blocks with Local Pre-ranking for Long Document Information Retrieval
• Graph-based Hierarchical Relevance Matching Signals for Ad-hoc Retrieval
• Generating Biomedical Question Answering Corpora from Q&A forums
• DeepTileBars: Visualizing Term Distribution for Neural Information Retrieval
• Machine Reading Comprehension for Answer Re-Ranking in Customer Support Chatbots
• RankNEAT: Outperforming Stochastic Gradient Search in Preference Learning Tasks
• Training Curricula for Open Domain Answer Re-Ranking
• Overcoming low-utility facets for complex answer retrieval
• Learning Colour Representations of Search Queries
• Separate and Attend in Personal Email Search
• GazBy: Gaze-Based BERT Model to Incorporate Human Attention in Neural Information Retrieval