Information Retrieval Model Research Articles

Knowledge enhanced edge-driven graph neural ranking for biomedical information retrieval

Neural networks used for information retrieval tend to capture textual matching signals between a query and a document. However, neural ranking models for biomedical information retrieval often struggle to semantically well match the query to the documents. The main reasons are that biomedical terms have many different representations and the fact description related to the query is non-consecutive and non-local in the documents. In this paper, we present an edge-driven graph neural ranking method for biomedical information retrieval by incorporating knowledge from medical databases. First, we propose to form an edge-driven graph by connecting some biomedical terms in the query and the document through different types of edges. Then, we design a novel way of knowledge integration to introduce knowledge related to biomedical terms into the graph and construct a knowledge-query-doc graph. Finally, a graph neural ranking model is applied to capture non-local and non-contiguous match signals between the query and the document. Experimental results show on the biomedical datasets that our method outperforms the advanced neural models. And further analysis shows that the knowledge integration method can well reduce the semantic gap between the query and the document, and our graph model can provide interpretation for matching between the query and the document.

Biomedical Information Retrieval with Positive-Unlabeled Learning and Knowledge Graphs

The rapid growth of biomedical publications has presented significant challenges in the field of information retrieval. Most existing work focuses on document retrieval given explicit queries. However, in real applications such as curated biomedical database maintenance, explicit queries are missing. In this paper, we propose a two-step model for biomedical information retrieval in the case that only a small set of example documents is available without explicit queries. Initially, we extract keywords from the observed documents using large pre-trained language models and biomedical knowledge graphs. These keywords are then enriched with domain-specific entities. Information retrieval techniques can subsequently use the collected entities to rank the documents. Following this, we introduce an iterative Positive-Unlabeled learning method to classify all unlabeled documents. Experiments conducted on the PubMed dataset demonstrate that the proposed technique outperforms the state-of-the-art positive-unlabeled learning methods. The results underscore the effectiveness of integrating large language models and biomedical knowledge graphs in improving zero-shot information retrieval performance in the biomedical domain.

Spectral clustering and query expansion using embeddings on the graph-based extension of the set-based information retrieval model

Spectral clustering and query expansion using embeddings on the graph-based extension of the set-based information retrieval model

A Study of Word Bigrams for Pseudo-relevance Feedback in Information Retrieval

Traditional information retrieval models mostly adopt a term independence assumption and are based on single terms or unigrams. Past efforts have attempted to go beyond this assumption, such as by using contiguous terms (i.e. word n-grams) or terms appearing in proximity. One such approach employs pseudo-relevance feedback (PRF) in an extended BM25 model, with an expanded query containing bigrams and proximity word pairs besides unigrams. However, the benefit of this approach over the traditional unigram PRF remains inconclusive. We speculate the uncertain effectiveness of bigram PRF in this past work is due to: (1) The new bigrams obtained from the expanded query may be formed by pairing unigrams drawn from different documents. These are potentially noise instead of relevant concepts; (2) The collection statistics of n-grams needed to calculate the document ranking functions, such as their document frequency, is not available in retrieval. Only estimates of these quantities are used instead. We suggest that these issues may be overcome by extracting word n-grams as single units in query expansion, and employing a document index that contains both unigrams and word n-grams. We demonstrate the approach for the case of bigram PRF in an extended BM25. Retrieval experiments are conducted on a range of standard test collections. For the majority of tested collections, the difference between values of the evaluation metrics (Mean Average Precision and the precision-oriented NDCG@20) obtained by our bigram PRF and the unigram PRF baseline is not statistically significant. Thus, our bigram PRF fails to improve over unigram PRF robustly across collections. An analysis of our results reveals ‘query drift’ due to bigram query expansion terms that represent too-broad topics as a cause for the failure of our approach.

Cross media knowledge information retrieval model based on D-S evidence theory

Cross media knowledge information retrieval provides strong support for information processing and utilization in the information society, but there are problems such as heterogeneity in cross media knowledge information. Therefore, a cross media knowledge information retrieval model using D-S evidence theory is proposed, which involves using approximate calculation methods to improve this theory for information fusion, reducing computational complexity, and using deep networks for fine-grained information retrieval to improve retrieval accuracy. The results showed that the improved theory enhanced computational efficiency by about 27.23 %. The memory usage was <60 %, and the average accuracy of information fusion reached 93.14 %. It also exhibited high recall and low false alarm rates. The cross media knowledge information retrieval model proposed in the study achieved accuracy of 92.64 %, 96.49 %, and 97.46 % on the three datasets used in the experiment, respectively. The study provides an effective, computationally efficient, and highly accurate model for cross media knowledge information retrieval, which is expected to promote research and application in this field. The combination of improved D-S evidence theory and deep networks provides a powerful approach to solving the problem of cross media heterogeneous information retrieval, which has a positive promoting effect on the processing and utilization of information in the information society.

A case study on decompounding in Indian language IR

Abstract Decompounding is an essential preprocessing step in text-processing tasks such as machine translation, speech recognition, and information retrieval (IR). Here, the IR issues are explored from five viewpoints. (A) Does word decompounding impact the Indian language IR? If yes, to what extent? (B) Can corpus-based decompounding models be used in the Indian language IR? If yes, how? (C) Can machine learning and deep learning-based decompounding models be applied in the Indian language IR? If yes, how? (D) Among the different decompounding models (corpus-based, hybrid machine learning-based, and deep learning-based), which provides the best effectiveness in the IR domain? (E) Among the different IR models, which provides the best effectiveness from the IR perspective? This study proposes different corpus-based, hybrid machine learning-based, and deep learning-based decompounding models in Indian languages (Marathi, Hindi, and Sanskrit). Moreover, we evaluate the effectiveness of each activity from an IR perspective only. It is observed that the different decompounding models improve IR effectiveness. The deep learning-based decompounding models outperform the corpus-based and hybrid machine learning-based models in Indian language IR. Among the different deep learning-based models, the Bi-LSTM-A model performs best and improves mean average precision (MAP) by 28.02% in Marathi. Similarly, the Bi-RNN-A model improves MAP by 18.18% and 6.1% in Hindi and Sanskrit, respectively. Among the retrieval models, the In_expC2 model outperforms others in Marathi and Hindi, and the BB2 model outperforms others in Sanskrit.

On Embedding Implementations in Text Ranking and Classification Employing Graphs

This paper aims to enhance the Graphical Set-based model (GSB) for ranking and classification tasks by incorporating node and word embeddings. The model integrates a textual graph representation with a set-based model for information retrieval. Initially, each document in a collection is transformed into a graph representation. The proposed enhancement involves augmenting the edges of these graphs with embeddings, which can be pretrained or generated using Word2Vec and GloVe models. Additionally, an alternative aspect of our proposed model consists of the Node2Vec embedding technique, which is applied to a graph created at the collection level through the extension of the set-based model, providing edges based on the graph’s structural information. Core decomposition is utilized as a method for pruning the graph. As a byproduct of our information retrieval model, we explore text classification techniques based on our approach. Node2Vec embeddings are generated by our graphs and are applied in order to represent the different documents in our collections that have undergone various preprocessing methods. We compare the graph-based embeddings with the Doc2Vec and Word2Vec representations to elaborate on whether our approach can be implemented on topic classification problems. For that reason, we then train popular classifiers on the document embeddings obtained from each model.

Revisiting Bag of Words Document Representations for Efficient Ranking with Transformers

Modern transformer-based information retrieval models achieve state-of-the-art performance across various benchmarks. The self-attention of the transformer models is a powerful mechanism to contextualize terms over the whole input but quickly becomes prohibitively expensive for long input as required in document retrieval. Instead of focusing on the model itself to improve efficiency, this paper explores different bag of words document representations that encode full documents by only a fraction of their characteristic terms, allowing us to control and reduce the input length. We experiment with various models for document retrieval on MS MARCO data, as well as zero-shot document retrieval on Robust04, and show large gains in efficiency while retaining reasonable effectiveness. Inference time efficiency gains are both lowering the time and memory complexity in a controllable way, allowing for further trading off memory footprint and query latency. More generally, this line of research connects traditional IR models with neural “NLP” models and offers novel ways to explore the space between (efficient, but less effective) traditional rankers and (effective, but less efficient) neural rankers elegantly.

Document Retrieval System for Biomedical Question Answering

In this paper, we describe our biomedical document retrieval system and answers extraction module, which is part of the biomedical question answering system. Approximately 26.5 million PubMed articles are indexed as a corpus with the Apache Lucene text search engine. Our proposed system consists of three parts. The first part is the question analysis module, which analyzes the question and enriches it with biomedical concepts related to its wording. The second part of the system is the document retrieval module. In this step, the proposed system is tested using different information retrieval models, like the Vector Space Model, Okapi BM25, and Query Likelihood. The third part is the document re-ranking module, which is responsible for re-arranging the documents retrieved in the previous step. For this study, we tested our proposed system with 6B training questions from the BioASQ challenge task. We obtained the best MAP score on the document retrieval phase when we used Query Likelihood with the Dirichlet Smoothing model. We used the sequential dependence model at the re-rank phase, but this model produced a worse MAP score than the previous phase. In similarity calculation, we included the Named Entity Recognition (NER), UMLS Concept Unique Identifiers (CUI), and UMLS Semantic Types of the words in the question to find the sentences containing the answer. Using this approach, we observed a performance enhancement of roughly 25% for the top 20 outcomes, surpassing another method employed in this study, which relies solely on textual similarity.

Application of Routing Algorithm Based on Logical Petri Nets in the Structural Design of Agricultural Products Logistics System

Abstract In agricultural services, logistics around agricultural products has been the focus of academic attention. To implement efficient and safe logistics of agricultural products and optimize logistics' overall efficiency and investment cost. This paper describes in detail the routing algorithm based on logical Petri nets, proposes a research system for agricultural product logistics based on the routing algorithm, and analyses system operation results through simulation experiments. The specific conclusions are as follows: (1) The logistics information modeling software based on logical Petri net is established, and the LPN model of the logistics information set is built based on it. (2) It is proposed that the server computer acts as a cloud service platform. The independent computer represents the traditional logistics information retrieval model, and the computational efficiency has been significantly improved. (3) The use of server computers to obtain logistics information in 300-400ms and the use of stand-alone computers to obtain logistics information in 50-120ms, the use of server computers to obtain logistics information is more efficient than using stand-alone computers. (4) Logistics data is a fast-growing trend; stand-alone computers cannot store massive data.

An Efficient Document Retrieval for Korean Open-Domain Question Answering Based on ColBERT

Open-domain question answering requires the task of retrieving documents with high relevance to the query from a large-scale corpus. Deep learning-based dense retrieval methods have become the primary approach for finding related documents. Although deep learning-based methods have improved search accuracy compared to traditional techniques, they simultaneously impose a considerable increase in computational burden. Consequently, research on efficient models and methods that optimize the trade-off between search accuracy and time to alleviate computational demands is required. In this paper, we propose a Korean document retrieval method utilizing ColBERT’s late interaction paradigm to efficiently calculate the relevance between questions and documents. For open-domain Korean question answering document retrieval, we construct a Korean dataset using various corpora from AI-Hub. We conduct experiments comparing the search accuracy and inference time among the traditional IR (information retrieval) model BM25, the dense retrieval approach utilizing BERT-based models for Korean, and our proposed method. The experimental results demonstrate that our approach achieves a higher accuracy than BM25 and requires less search time than the dense retrieval method employing KoBERT. Moreover, the most outstanding performance is observed when using KoSBERT, a pre-trained Korean language model that learned to position semantically similar sentences closely in vector space.

Evaluating Semantic and Personalized Information Retrieval with Deep Learning Models: A Performance Metrics Analysis

Evaluation of the Information Retrieval (IR) model is an important, complex process that involves evaluating many aspects of performance using quantitative measurements called evaluation metrics. This work explores the efficacy of Deep learning (DL) models in personalized and semantic information retrieval (SIR), an area that has witnessed significant advancements in the last few years. A thorough analysis of several DL models for search engine index retrieval is provided. The particular task, relevant standards, and desirable qualities, among others, influence the use of evaluation metrics in IR. A comprehensive evaluation of the performance of an IR system uses several metrics. The paper emphasizes how DL models can produce IR results that are more accurate, with an emphasis on choosing important metrics for evaluation. Among these metrics, on a given input query, a recall of 1.0 indicates full coverage in detecting pertinent cases, while a precision of 0.5 indicates 50% accuracy in positive predictions on a given input query. At 0.6667, the F1-score strikes a compromise between recall and precision. The model's ability to rank and prioritize is assessed by Mean Average Precision (MAP) at 0.5426 and Normalized Discounted Cumulative Gain (NDCG) at 0.8896. Precision in the top 5 suggestions is evaluated by Precision@5, which has a value of 0.6. This is important in situations where user attention or recommendation space are restricted. These measurements offer priceless information on the efficacy of IR models.

Research on privacy information retrieval model based on hybrid homomorphic encryption

The computational complexity of privacy information retrieval protocols is often linearly related to database size. When the database size is large, the efficiency of privacy information retrieval protocols is relatively low. This paper designs an effective privacy information retrieval model based on hybrid fully homomorphic encryption. The assignment method is cleverly used to replace a large number of homomorphic encryption operations. At the same time, the multiplicative homomorphic encryption scheme is first used to deal with the large-scale serialization in the search, and then the fully homomorphic encryption scheme is used to deal with the remaining simple operations. The depth of operations supported by the fully homomorphic scheme no longer depends on the size of the database, but only needs to support the single homomorphic encryption scheme to decrypt the circuit depth. Based on this hybrid homomorphic encryption retrieval model, the efficiency of homomorphic privacy information retrieval model can be greatly improved.

NIR-Prompt: A Multi-task Generalized Neural Information Retrieval Training Framework

Information retrieval aims to find information that meets users’ needs from the corpus. Different needs correspond to different IR tasks such as document retrieval, open-domain question answering, retrieval-based dialogue, and so on, while they share the same schema to estimate the relationship between texts. It indicates that a good IR model can generalize to different tasks and domains. However, previous studies indicate that state-of-the-art neural information retrieval (NIR) models, e.g., pre-trained language models (PLMs) are hard to generalize. It is mainly because the end-to-end fine-tuning paradigm makes the model overemphasize task-specific signals and domain biases but loses the ability to capture generalized essential signals. To address this problem, we propose a novel NIR training framework named NIR-Prompt for retrieval and reranking stages based on the idea of decoupling signal capturing and combination. NIR-Prompt exploits Essential Matching Module (EMM) to capture the essential matching signals and gets the description of tasks by Matching Description Module (MDM). The description is used as task-adaptation information to combine the essential matching signals to adapt to different tasks. Experiments under in-domain multi-task, out-of-domain multi-task, and new task adaptation settings show that NIR-Prompt can improve the generalization of PLMs in NIR for both retrieval and reranking stages compared with baselines.

Conditional variational autoencoder for query expansion in ad-hoc information retrieval

Query expansion (QE) is commonly used to improve the performance of traditional information retrieval (IR) models. With the adoption of deep learning in IR research, neural QE models have emerged in recent years. Many of these models focus on learning embeddings by leveraging query-document relevance. These embedding models allow computing semantic similarities between queries and documents to generate expansion terms. However, existing models often ignore query-document interactions. This research aims to address that gap by proposing a QE model using a conditional variational autoencoder. It first maps a query-document pair into a latent space based on their interaction, then estimates an expansion model from that latent space. The proposed model is trained on relevance feedback data and generates expansions using pseudo-relevance feedback at test time. The proposed model is evaluated on three standard TREC collections for document ranking: AP and Robust 04 and GOV02, and the MS MARCO dataset for passage ranking. Results show the model outperforms state-of-the-art traditional and neural QE models. It also demonstrates higher additivity with neural matching than baselines.

Cross Language Information Retrieval based on Automatic Query Translation for Marathi Documents

Abstract: The current research article explores the realm of Cross Language Information Retrieval (CLIR) and its significance in the digital age. It addresses the challenges faced in CLIR, including lexical and semantic disparities, the scarcity of parallel corpora, cultural nuances, and more. The article discusses innovative solutions encompassing Machine Translation, Query Expansion, Cross-Lingual Word Embeddings, and Multilingual Information Retrieval Models to enhance CLIR's effectiveness. Furthermore, it sheds light on Information Retrieval Models, such as the Boolean Model, Vector Space Model (VSM), and Probabilistic Models, explaining their principles and applications. The study also presents experimental results highlighting the limitations of monolingual IR models and the effectiveness of crosslingual techniques, such as translation and query expansion, in improving CLIR, making it a valuable tool for accessing information across languages.

Enhancing learning and exploratory search with concept semantics in online healthcare knowledge management systems: An interactive knowledge visualization approach

Healthcare is a knowledge intensive process, which requires large amount of information from various healthcare knowledge providers. However, this growing information makes challengeable for users in finding useful healthcare information to support their healthcare decision making. Knowledge management and knowledge management systems are drawing increasing attention in healthcare community. This paper proposes an online healthcare knowledge management framework and system prototype, featured with integrating knowledge learning and exploratory search for solving problems efficiently. Inspired by the knowledge to action framework in healthcare and the iterative process in software engineering, an iterative and incremental knowledge management model is designed to organize the activities during the knowledge learning and searching procedure. Supported by the knowledge-enhanced infrastructure of online healthcare information designed, domain knowledge learning and task knowledge learning environment can work collaboratively and interactively to improve the result of learning and exploratory search. Knowledge base also provides helpful information in an interactive way for solving problems. Given exploiting probabilistic information retrieval model and concept semantics described in Description Logics, a search intent discovery method and search intent guided learning and searching mechanism are designed within this environment for achieving the learning and search goal. The results of the experiments have proved the efficacy of the methods and process model used in the leaning-focused healthcare knowledge management system.

Towards Robust Neural Rankers with Large Language Model: A Contrastive Training Approach

Pre-trained language model-based neural rankers have been widely applied in information retrieval (IR). However, the robustness issue of current IR models has not received sufficient attention, which could significantly impact the user experience in practical applications. In this study, we focus on the defensive ability of IR models against query attacks while guaranteeing their retrieval performance. We discover that improving the robustness of IR models not only requires a focus on model architecture and training methods but is also closely related to the quality of data. Different from previous research, we use large language models (LLMs) to generate query variations with the same intent, which exhibit richer and more realistic expressions while maintaining consistent query intent. Based on LLM-generated query variations, we propose a novel contrastive training framework that substantially enhances the robustness of IR models to query perturbations. Specifically, we combine the contrastive loss in the representation space of query variations with the ranking loss in the retrieval training stage to improve the model’s ability to understand the underlying semantic information of queries. Experimental results on two public datasets, WikiQA and ANTIQUE, demonstrate that the proposed contrastive training approach effectively improves the robustness of models facing query attack scenarios while outperforming baselines in retrieval performance. Compared with the best baseline approach, the improvements in average robustness performance of Reranker IR models are 24.9%, 26.5%, 27.0%, and 75.0% on WikiQA and 8.7%, 1.9%, 6.3%, and 13.6% on ANTIQUE, in terms of the MAP (Mean Average Precision), MRR (Mean Reciprocal Rank), nDCG@10 (Normalized Discounted Cumulative Gain) and P@10 (Precision), respectively.

Connectomes and properties of quantum entanglement

Topological quantum field theories (TQFT) encode properties of quantum states in the topological features of abstract manifolds. One can use the topological avatars of quantum states to develop intuition about different concepts and phenomena of quantum mechanics. In this paper we focus on the class of simplest topologies provided by a specific TQFT and investigate what the corresponding states teach us about entanglement. These “planar connectome” states are defined by graphs of simplest topology for a given adjacency matrix. In the case of bipartite systems the connectomes classify different types of entanglement matching the classification of stochastic local operations and classical communication (SLOCC). The topological realization makes explicit the nature of entanglement as a resource and makes apparent a number of its properties, including monogamy and characteristic inequalities for the entanglement entropy. It also provides tools and hints to engineer new measures of entanglement and other applications. Here the approach is used to construct purely topological versions of the dense coding and quantum teleportation protocols, giving diagrammatic interpretation of the role of entanglement in quantum computation and communication. Finally, the topological concepts of entanglement and quantum teleportation are employed in a simple model of information retrieval from a causally disconnected region, similar to the interior of an evaporating black hole.

Explainable online health information truthfulness in Consumer Health Search.

People are today increasingly relying on health information they find online to make decisions that may impact both their physical and mental wellbeing. Therefore, there is a growing need for systems that can assess the truthfulness of such health information. Most of the current literature solutions use machine learning or knowledge-based approaches treating the problem as a binary classification task, discriminating between correct information and misinformation. Such solutions present several problems with regard to user decision making, among which: (i) the binary classification task provides users with just two predetermined possibilities with respect to the truthfulness of the information, which users should take for granted; indeed, (ii) the processes by which the results were obtained are often opaque and the results themselves have little or no interpretation. To address these issues, we approach the problem as an ad hoc retrieval task rather than a classification task, with reference, in particular, to the Consumer Health Search task. To do this, a previously proposed Information Retrieval model, which considers information truthfulness as a dimension of relevance, is used to obtain a ranked list of both topically-relevant and truthful documents. The novelty of this work concerns the extension of such a model with a solution for the explainability of the results obtained, by relying on a knowledge base consisting of scientific evidence in the form of medical journal articles. We evaluate the proposed solution both quantitatively, as a standard classification task, and qualitatively, through a user study to examine the "explained" ranked list of documents. The results obtained illustrate the solution's effectiveness and usefulness in making the retrieved results more interpretable by Consumer Health Searchers, both with respect to topical relevance and truthfulness.