A Deep Look into Neural Ranking Models for Information Retrieval

By Jiafeng Guo, Yixing Fan, Liang Pang, Liu Yang, Qingyao Ai, Hamed Zamani, Chen Wu, W. Bruce Croft, Xueqi Cheng. 2019

386 citations recorded by the source corpus.

Abstract Ranking models lie at the heart of research on information retrieval (IR). During the past decades, different techniques have been proposed for constructing ranking models, from traditional heuristic methods, probabilistic methods, to modern machine learning methods. Recently, with the advance of deep learning technology, we have witnessed a growing body of work in applying shallow or deep neural networks to the ranking problem in IR, referred to as neural ranking models in this paper. The power of neural ranking models lies in the ability to learn from the raw text inputs for the ranking problem to avoid many limitations of hand-crafted features. Neural networks have sufficient capacity to model complicated tasks, which is needed to handle the complexity of relevance estimation in ranking. Since there have been a large variety of neural ranking models proposed, we believe it is the right time to summarize the current status, learn from existing methodologies, and gain some insights for future development. In contrast to existing reviews, in this survey, we will take a deep look into the neural ranking models from different dimensions to analyze their underlying assumptions, major design principles, and learning strategies. We compare these models through benchmark tasks to obtain a comprehensive empirical understanding of the existing techniques. We will also discuss what is missing in the current literature and what are the promising and desired future directions.

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

ColBERT: Efficient and Effective Passage Search via Contextualized Late Interaction over BERT (2020, methodology): “, Figure 2 (b) and (c)) tend to be superior for IR tasks [8, 21], a representation-focused model—by isolating the computations among q and d—makes it possible to precompute document representations offline [41], greatly reducing the computational load per query.”
Simple Applications of BERT for Ad Hoc Document Retrieval (2019, background): “4 As a point of comparison, in the most recent survey of neural ranking models by Guo et al. (2019), the best AP on Robust04 is in the 0 .”
Context-Aware Document Term Weighting for Ad-Hoc Search (2020, background): “This is a common issue in controlled vocabularies [35] and representation-based neural ranking models [16, 17].”
Multi-step Entity-centric Information Retrieval for Multi-Hop Question Answering (2019, neutral): “Neural ranking models have shown great potential and have been widely adopted in the IR community (Dehghani et al., 2017; Guo et al., 2019; Mitra et al., 2017; Zamani et al., 2018, inter-alia).”
Long Document Ranking with Query-Directed Sparse Transformer (2020, background): “Neural models have demonstrated significant advances across various ranking tasks (Guo et al., 2019).”
The Fact Extraction and VERification Over Unstructured and Structured information (FEVEROUS) Shared Task (2021, methodology): “Given that neural ranking methods have shown success in the IR community (Guo et al., 2020), we used one as part of our extension to the baseline, where a re-ranker provides a score for every (query, table) pair.”
RepBERT: Contextualized Text Embeddings for First-Stage Retrieval (2020, methodology): “Because our model can be categorized as representation-focused models [9] in IR community, we call the proposed model RepBERT .”
Dense Text Retrieval Based on Pretrained Language Models: A Survey (2022, background): “We are aware of several closely related surveys or books on dense retrieval [24]–[26], [29], [35].”
Beyond 512 Tokens: Siamese Multi-depth Transformer-based Hierarchical Encoder for Document Matching (2020, background): “ber of neural matching models havebeenproposedforinformationretrieval[7,11,12,18,19,32–34, 39]. These models can be classified into therepresentation-focused models and the interaction-focused models [7, 8]. The representationfocused models learn the representations of queries and documents separate”
DiPair: Fast and Accurate Distillation for Trillion-ScaleText Matching and Pair Modeling (2020, background): “For example, modeling the relationship between queries and documents / ad keywords is central to search engines / digital advertisement systems (Li and Xu, 2014; Guo et al., 2019).”
Conformer-Kernel with Query Term Independence for Document Retrieval (2020, methodology): “Efficient retrieval using complex machine learned relevance functions is an important challenge in neural IR [Mitra and Craswell, 2018, Guo et al., 2019].”
An Analysis of BERT in Document Ranking (2020, neutral): “[11] designed a representation-focused [6] BERT ranker.”
Deep Learning for Matching in Search and Recommendation (2018, background): “Both traditional machine learning approaches (Li and Xu, 2014) and deep learning approaches (Guo et al., 2019b; Mitra and Craswell, 2018; Onal et al., 2018) have been developed for semantic matching.”
Selective Weak Supervision for Neural Information Retrieval (2020, background): “Conv-KNRM and BERT often require large scale relevance supervision [14, 47].”
Local Self-Attention over Long Text for Efficient Document Retrieval (2020, background): “Neural models have shown successful results in a number of IR tasks [9, 10, 17].”
A Survey of Model Architectures in Information Retrieval (2025, unknown): “The model constructs matching histograms for each query term to capture the distribution of matching signals across document terms.”
LocAgent: Graph-Guided LLM Agents for Code Localization (2025, unknown): “Unlike traditional retrieval tasks that primarily focus on lexical or semantic matching between queries and documents (Guo et al., 2016, 2020), code localization requires bridging the gap between natural language and programming languages.”