Traditional Information Retrieval Research Articles

FMCF: A fusing multiple code features approach based on Transformer for Solidity smart contracts source code summarization

A smart contract is a software program executed on a blockchain, designed to facilitate functionalities such as contract execution, asset administration, and identity validation within a secure and decentralized ecosystem. Summarizing the code of Solidity smart contracts aids developers in promptly grasping essential functionalities, thereby enhancing the security posture of Ethereum-based projects. Existing smart contract code summarization works mainly use traditional information retrieval and single code features, resulting in suboptimal performance. In this study, we propose a fusing multiple code features (FMCF) approach based on Transformer for Solidity summarization. First, FMCF created contract integrity modeling and state immutability modeling in the data preprocessing stage to process and filter data that meets security conditions. At the same time, FMCF retains the self-attention mechanism to construct the Graph Attention Network (GAT) encoder and CodeBERT encoder, which respectively extract multiple feature vectors of the code to ensure the integrity of the source code information. Furthermore, the FMCF uses a weighted summation method to input these two types of feature vectors into the feature fusion module for fusion and inputs the fused feature vectors into the Transformer decoder to obtain the final smart contract code summarization. The experimental results show that FMCF outperforms the standard baseline methods by 12.45% in the BLEU score and maximally preserves the semantic information and syntax structures of the source code. The results demonstrate that the FMCF can provide a good direction for future research on smart contract code summarization, thereby helping developers enhance the security of development projects.

A question-answering framework for automated abstract screening using large language models.

This paper aims to address the challenges in abstract screening within systematic reviews (SR) by leveraging the zero-shot capabilities of large language models (LLMs). We employ LLM to prioritize candidate studies by aligning abstracts with the selection criteria outlined in an SR protocol. Abstract screening was transformed into a novel question-answering (QA) framework, treating each selection criterion as a question addressed by LLM. The framework involves breaking down the selection criteria into multiple questions, properly prompting LLM to answer each question, scoring and re-ranking each answer, and combining the responses to make nuanced inclusion or exclusion decisions. Large-scale validation was performed on the benchmark of CLEF eHealth 2019 Task 2: Technology-Assisted Reviews in Empirical Medicine. Focusing on GPT-3.5 as a case study, the proposed QA framework consistently exhibited a clear advantage over traditional information retrieval approaches and bespoke BERT-family models that were fine-tuned for prioritizing candidate studies (ie, from the BERT to PubMedBERT) across 31 datasets of 4 categories of SRs, underscoring their high potential in facilitating abstract screening. The experiments also showcased the viability of using selection criteria as a query for reference prioritization. The experiments also showcased the viability of the framework using different LLMs. Investigation justified the indispensable value of leveraging selection criteria to improve the performance of automated abstract screening. LLMs demonstrated proficiency in prioritizing candidate studies for abstract screening using the proposed QA framework. Significant performance improvements were obtained by re-ranking answers using the semantic alignment between abstracts and selection criteria. This further highlighted the pertinence of utilizing selection criteria to enhance abstract screening.

Enhancing Medical Image Retrieval with UMLS-Integrated CNN-Based Text Indexing.

In recent years, Convolutional Neural Network (CNN) models have demonstrated notable advancements in various domains such as image classification and Natural Language Processing (NLP). Despite their success in image classification tasks, their potential impact on medical image retrieval, particularly in text-based medical image retrieval (TBMIR) tasks, has not yet been fully realized. This could be attributed to the complexity of the ranking process, as there is ambiguity in treating TBMIR as an image retrieval task rather than a traditional information retrieval or NLP task. To address this gap, our paper proposes a novel approach to re-ranking medical images using a Deep Matching Model (DMM) and Medical-Dependent Features (MDF). These features incorporate categorical attributes such as medical terminologies and imaging modalities. Specifically, our DMM aims to generate effective representations for query and image metadata using a personalized CNN, facilitating matching between these representations. By using MDF, a semantic similarity matrix based on Unified Medical Language System (UMLS) meta-thesaurus, and a set of personalized filters taking into account some ranking features, our deep matching model can effectively consider the TBMIR task as an image retrieval task, as previously mentioned. To evaluate our approach, we performed experiments on the medical ImageCLEF datasets from 2009 to 2012. The experimental results show that the proposed model significantly enhances image retrieval performance compared to the baseline and state-of-the-art approaches.

Next generation of electronic medical record search engines to support chart reviews: A systematic user study and future research direction

ObjectiveLittle research has been done on the user-centered document ranking approach, especially in a crowdsourcing chart review environment. As the starting point of designing and implementing the next generation of Electronic Medical Record (EMR) search engines, a systematic user study is needed to better understand the users' needs, challenges, and future research directions of EMR search engines. Materials and methodsOne primary observation during the user study is the need for a ranking method to better support the so-called "early stopping" reviewing strategy (i.e., reviewing only a subset of EMRs of one patient to make the final decision) during the clinical chart reviews. The authors proposed two novel user-centered ranking metrics: "critical documents" and "negative guarantee ratio," to better measure the power of a ranking method in supporting the “early stopping” requirements during clinical chart reviews. ResultsThe evaluation results show that i) traditional information retrieval metrics, such as the precision-at-K, have limitations in guiding the design and development of EMR search engines to better support clinical chart reviews; ii) there is not a global optimal ranking method that fits the needs of different chart reviews and different users; iii) a learning-to-rank approach cannot guarantee a stable and optimal ranking for different chart reviews and different users; and iv) A user-centered ranking metric, such as the negative guarantee ratio (NGR) metric is able to measure the “early-stopping” performance of ranking methods. ConclusionsUser-centered ranking metrics can better measure the power of ranking methods in supporting clinical chart reviews. Future research should explore more user-centered ranking metrics and evaluate their impact on real-world EMR search engines.

Towards a Regulatory Framework for Electronic Medical Record Data Visualization.

Physicians struggle to retrieve data from electronic medical records. We evaluated a digital tool that enhances physician efficiency in retrieving and analyzing patient information for treatment decision-making. Our use case is the care of diabetic patients. Evaluation results showed that healthcare providers who used the i4C (Insights for Care) dashboard experienced greater time efficiency than those who used traditional EMR information retrieval methods. A comprehensive evaluation of the i4C Dashboard confirms its effectiveness in facilitating diabetic care data management, as well as its potential application to a wide range of healthcare scenarios. In order to further maximize its effectiveness on clinical efficiency and patient care, future research should focus on improving its usability and scalability.

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.

Content-based image retrieval based on corel dataset using deep learning

A popular technique for retrieving images from huge and unlabeled image databases are content-based-image-retrieval (CBIR). However, the traditional information retrieval techniques do not satisfy users in terms of time consumption and accuracy. Additionally, the number of images accessible to users are growing due to web development and transmission networks. As the result, huge digital image creation occurs in many places. Therefore, quick access to these huge image databases and retrieving images like a query image from these huge image collections provides significant challenges and the need for an effective technique. Feature extraction and similarity measurement are important for the performance of a CBIR technique. This work proposes a simple but efficient deep-learning framework based on convolutional-neural networks (CNN) for the feature extraction phase in CBIR. The proposed CNN aims to reduce the semantic gap between low-level and high-level features. The similarity measurements are used to compute the distance between the query and database image features. When retrieving the first 10 pictures, an experiment on the Corel-1K dataset showed that the average precision was 0.88 with Euclidean distance, which was a big step up from the state-of-the-art approaches.

Rethinking E-Commerce Search

E-commerce search and recommendation usually operate on structured data such as product catalogs and taxonomies. However, creating better search and recommendation systems often requires a large variety of unstructured data including customer reviews and articles on the web. Traditionally, the solution has always been converting unstructured data into structured data through information extraction, and conducting search over the structured data. However, this is a costly approach that often has low quality. In this paper, we envision a solution that does entirely the opposite. Instead of converting unstructured data (web pages, customer reviews, etc) to structured data, we instead convert structured data (product inventory, catalogs, taxonomies, etc) into textual data, which can be easily integrated into the text corpus that trains LLMs. Then, search and recommendation can be performed through a Q/A mechanism through an LLM instead of using traditional information retrieval methods over structured data.

Towards a Method to Enable the Selection of Physical Models within the Systems Engineering Process: A Case Study with Simulink Models

The use of different techniques and tools is a common practice to cover all stages in the development life-cycle of systems generating a significant number of work products. These artefacts are frequently encoded using diverse formats, and often require access through non-standard protocols and formats. In this context, Model-Based Systems Engineering (MBSE) emerges as a methodology to shift the paradigm of Systems Engineering practice from a document-oriented environment to a model-intensive environment. To achieve this major goal, a formalised application of modelling is employed throughout the life-cycle of systems to generate various system artefacts represented as models, such as requirements, logical models, and multi-physics models. However, the mere use of models does not guarantee one of the main challenges in the Systems Engineering discipline, namely, the reuse of system artefacts. Considering the fact that models are becoming the main type of system artefact, it is necessary to provide the capability to properly and efficiently represent and retrieve the generated models. In light of this, traditional information retrieval techniques have been widely studied to match existing software assets according to a set of capabilities or restrictions. However, there is much more at stake than the simple retrieval of models or even any piece of knowledge. An environment for model reuse must provide the proper mechanisms to (1) represent any piece of data, information, or knowledge under a common and shared data model, and (2) provide advanced retrieval mechanisms to elevate the meaning of information resources from text-based descriptions to concept-based ones. This need has led to novel methods using word embeddings and vector-based representations to semantically encode information. Such methods are applied to encode the information of physical models while preserving their underlying semantics. In this study, a text corpus from MATLAB Simulink models was preprocessed using Natural Language Processing (NLP) techniques and trained to generate word vector representations. Then, the presented method was validated using a testbed of MATLAB Simulink physical models in which verbalisations of models are transformed into vectors. The effectiveness of the proposed solution was assessed through a use case study. Evaluation of the results demonstrates a precision value of 0.925, a recall value of 0.865, and an F1 score of 0.884.

Enhancing Multimodal Information Retrieval Through Integrating Data Mining and Deep Learning Techniques

Multimodal information retrieval, the task of re trieving relevant information from heterogeneous data sources such as text, images, and videos, has gained significant attention in recent years due to the proliferation of multimedia content on the internet. This paper proposes an approach to enhance multimodal information retrieval by integrating data mining and deep learning techniques. Traditional information retrieval systems often struggle to effectively handle multimodal data due to the inherent complexity and diversity of such data sources. In this study, we leverage data mining techniques to preprocess and structure multimodal data efficiently. Data mining methods enable us to extract valuable patterns, relationships, and features from different modalities, providing a solid foundation for sub- sequent retrieval tasks. To further enhance the performance of multimodal information retrieval, deep learning techniques are employed. Deep neural networks have demonstrated their effectiveness in various multimedia tasks, including image recognition, natural language processing, and video analysis. By integrating deep learning models into our retrieval framework, we aim to capture complex intermodal dependencies and semantically rich representations, enabling more accurate and context-aware retrieval.

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.

A template-based approach for question answering over knowledge bases

In this paper, we address the problem of answering complex questions formulated by users in natural language. Since traditional information retrieval systems are not suitable for complex questions, these questions are usually run over knowledge bases, such as Wikidata or DBpedia. We propose a semi-automatic approach for transforming a natural language question into a SPARQL query that can be easily processed over a knowledge base. The approach applies classification techniques to associate a natural language question with a proper query template from a set of predefined templates. The nature of our approach is semi-automatic as the query templates are manually written by human assessors, who are the experts of the knowledge bases, whereas the classification and query processing steps are completely automatic. Our experiments on the large-scale CSQA dataset for question-answering corroborate the effectiveness of our approach.

SPRF: A semantic Pseudo-relevance Feedback enhancement for information retrieval via ConceptNet

Pseudo-relevance feedback is a widely acclaimed technique for information retrieval. However, traditional information retrieval approaches typically process the original query into individual terms, often overlooking the consideration of the semantic information in the query term itself, by instead focusing on features such as term frequency and inverse document frequency. In this paper, we propose a new semantic-based Pseudo-relevance Feedback model (SPRF) based on the PRF framework. Our SPRF model leverages ConceptNet to provide comprehensive semantic information between terms. It not only considers the query’s importance in the collection but also integrates its semantic information into the PRF framework to enhance the selection of query expansion terms, leading to more precise feedback documents for users. A series of experimental results show that our proposed SPRF model is feasible. Our model achieves good performance in terms of the MAP, P@10, NDCG and MRR metrics and demonstrates advantages over the baseline models, the state-of-the-art models and several neural network-based methods. By comparing and analyzing the methods in a sample case, the extended terms resulting from the proposed model are shown to be in better semantic agreement with the given query.

An efficient ontology model with query execution for accurate document content extraction

<span lang="EN-US">The technique of extracting important documents from massive data collections is known as information retrieval (IR). The dataset provider coupled with the increasing demand for high-quality retrieval results, has resulted in traditional information retrieval approaches being increasingly insufficient to meet the challenge of providing high-quality search results. Research has concentrated on information retrieval and interactive query formation through ontologies in order to overcome these challenges, with a specific emphasis on enhancing the functionality between information and search queries in order to bring the outcome sets closer to the research requirements of users. In the context of document retrieval technologies, it is a process that assists researchers in extracting documents from data collections. It is discussed in this research how to use ontology-based information retrieval approaches and techniques, taking into account the issues of ontology modelling, processing, and the transformation of ontological knowledge into database search queries. In this research work, an efficient optimized ontology model with query execution for content extraction from documents (OOM-QE-CE) is proposed. The existing ontology-to-database transformation and mapping methodologies are also thoroughly examined in terms of data and semantic loss, structural mapping and domain knowledge applicability.</span>

Ontology-based hierarchical retrieval model for digital English teaching information

In order to overcome the low recall and precision of traditional English teaching information retrieval model, this paper designs a hierarchical retrieval model of digital English teaching information based on ontology. TF-IDF and simhash algorithm are used to judge the similarity of digital English teaching database documents and calculate the weight of English teaching information retrieval keywords. Using the relationship between different retrieval concepts to build a semantic network diagram, a hierarchical retrieval model of digital English teaching information was built, the retrieval results according to the user's interest were adjusted, and more accurate retrieval results are obtained. The experimental results show that the recall rate of the model is more than 94%, the precision rate is more than 96%, and the average retrieval time is only 0.44s, which shows that the recall rate and precision rate of the design model are higher, and the retrieval time is shorter.

Enhancing Semantic Code Search with Deep Graph Matching

The job of discovering appropriate code snippets against a natural language query is an important task for software developers. Appropriate code retrieval increases software productivity and quality as well. In contrast to traditional information retrieval techniques, code search necessitates bridging the semantic breach between programming languages and natural language to search code fragments. Deep neural networks for search codes have recently been a hot topic in research. The standard neural code quest approaches present source code and query in the form of text as independent embedding, then calculate the semantic similarity between them using vector distance (e.g., using cosine similarity). Although recent research utilized query and code snippets during code search, it overlooked the contained rich semantic information and deep structural features between them. In this study, we are also dealing with the problem of code search by providing a deep neural solution that facilitates software developers during software development. Our proposed model effectively used neural graph matching and a searching approach for semantic code retrieval. It first converts both query and code fragments in graph format and then the semantic matching module is used to facilitate the process of matching that will retrieve the best-matched code snippets. It not only exploits the enriched semantic meanings and features, but it also uses the cross-attention mechanism to learn the fine-grained similarity that exists between query and code. The proposed model’s evaluation is done using the Codesearchnet dataset with six representative programming languages. It provides comparatively good results as compared to existing baselines. It enables users to find required code snippets, and ranking is used to retrieve top 10 results. The accuracy of the proposed system is approximately 97%.

NLP-Based Bi-Directional Recommendation System: Towards Recommending Jobs to Job Seekers and Resumes to Recruiters

For more than ten years, online job boards have provided their services to both job seekers and employers who want to hire potential candidates. The provided services are generally based on traditional information retrieval techniques, which may not be appropriate for both job seekers and employers. The reason is that the number of produced results for job seekers may be enormous. Therefore, they are required to spend time reading and reviewing their finding criteria. Reciprocally, recruitment is a crucial process for every organization. Identifying potential candidates and matching them with job offers requires a wide range of expertise and knowledge. This article proposes a reciprocal recommendation based on bi-directional correspondence as a way to support both recruiters’ and job seekers’ work. Recruiters can find the best-fit candidates for every job position in their job postings, and job seekers can find the best-match jobs to match their resumes. We show how machine learning can solve problems in natural language processing of text content and similarity scores depending on job offers in major Saudi cities scraped from Indeed. For bi-directional matching, a similarity calculation based on the integration of explicit and implicit job information from two sides (recruiters and job seekers) has been used. The proposed system is evaluated using a resume/job offer dataset. The performance of generated recommendations is evaluated using decision support measures. Obtained results confirm that the proposed system can not only solve the problem of bi-directional recommendation, but also improve the prediction accuracy.

Intelligent information retrieval based on Ontology

Traditional search engines can no longer meet the increasing amount of information in today's world, and often face problems such as blurred results and long search times for some queries. And the addition of ontology makes information retrieval more intelligent, so that information data between various fields can be shared, to achieve the purpose of getting the answers users want quickly and accurately. This paper firstly introduces the definition of ontology and analyzes the traditional information retrieval system, and finally summarizes some experimental results in recent years in ontology-based intelligent information retrieval and reviews them to get a more efficient and accurate information retrieval system to achieve the effect of improving the accuracy of search engines.

Analyzing the past, improving the future: a multiscale opinion tracking model for optimizing business performance

The complexity of business decision-making has increased over the years. It is essential for managers to gain a confident understanding of their business environments in order to make successful decisions. With the growth of opinion-rich web resources such as social media, discussion forums, review sites, news corpora, and blogs available on the internet, product and service reviews have become an essential source of information. In a data-driven world, they will improve services and operational insights to achieve real business benefits and help enterprises remain competitive. Despite the prevalence of textual data, few studies have demonstrated the effectiveness of real-time text mining and reporting tools in firms and organizations. To address this aspect of decision-making, we have developed and evaluated an unsupervised learning system to automatically extract and classify topics and their emotion score in text streams. Data were collected from commercial websites, open-access databases, and social networks to train the model. In the experiment, the polarity score was quantified at four different levels: word, sentence, paragraph, and the entire text using Latent Dirichlet Allocation (LDA). Using subjective data mining, we demonstrate how to extract, summarize, and track various aspects of information from the Web and help traditional information retrieval (IR) systems to capture more information. An opinion tracking system presented by our model extracts subjective information, classifies them, and tracks opinions by utilizing location, time, and reviewers’ positions. Using the online-offline data collection technique, we can update the library topic in real-time to provide users with a market opinion tracker. For marketing or economic research, this approach may be useful. In the experiment, the new model is applied to a case study to demonstrate how the business process improves.

LawRec: Automatic Recommendation of Legal Provisions Based on Legal Text Analysis.

Smart court technologies are making full use of modern science to promote the modernization of the trial system and trial capabilities, for example, artificial intelligence, Internet of things, and cloud computing. The smart court technologies can improve the efficiency of case handling and achieving convenience for the people. Article recommendation is an important part of intelligent trial. For ordinary people without legal background, the traditional information retrieval system that searches laws and regulations based on keywords is not applicable because they do not have the ability to extract professional legal vocabulary from complex case processes. This paper proposes a law recommendation framework, called LawRec, based on Bidirectional Encoder Representation from Transformers (BERT) and Skip-Recurrent Neural Network (Skip-RNN) models. It intends to integrate the knowledge of legal provisions with the case description and uses the BERT model to learn the case description text and legal knowledge, respectively. At last, laws and regulations for cases can be recommended. Experiment results show that the proposed LawRec can achieve better performance than state-of-the-art methods.