Knowledge Retrieval

Retrieval-Augmented Generation (RAG) architectures represent a hybrid approach that blends information retrieval with generative modeling to tackle complex natural language processing (NLP) tasks. A key challenge in these systems is optimizing the balance between retrieval depth and generative coherence. Retrieval depth refers to the number of documents retrieved and utilized by the generative model, while generative coherence is the degree to which the generated output is relevant, contextually accurate, and logically consistent with the retrieved information. This paper proposes the RAG Optimization Framework (ROF), designed to fine-tune these factors and enhance performance across diverse applications. We examine various strategies to adjust retrieval depth dynamically, ensuring relevant data retrieval, and we explore techniques to maintain coherence in generative outputs. In addition, this paper investigates how multi-step retrieval can improve performance by progressively refining the information provided to the model. This framework's applications in fields like healthcare and financial document analysis are also discussed, illustrating its potential to significantly enhance RAG systems.

This study investigates the relationship between knowledge management and employee effectiveness of deposit money banks in Nigeria using cross-sectional survey. Sixty branch managers from ten deposit money banks operating in Abakaliki, Ebonyi State were surveyed using convenience sampling technique. Forty four copies of questionnaire were filled Statistical tool used to analyse the hypotheses is Pearson Product Moment Correlation Coefficient (rs). Respondents’ profiles were analysed using descriptive statistics. The study found that knowledge management has positive significant association with employee effectiveness in deposit money banks in Abakaliki. It concluded that knowledge management measured in terms of knowledge acquisition, knowledge dissemination, knowledge conservation and knowledge retrieval enhances employee effectiveness when resources are provided for the employees quite apart from setting goals on how jobs should be carried out.  Based on the conclusion above, one ...

Web 2.0 has changed the technological landscape of the Internet computing world today. The shift from traditional web which is also known as Web 1.0 is forced by the growing need for more efficient information sharing, collaboration, and business processes. The disclosure of personal/organizational information in Web 2.0 via social networks, digital contributions and data feeds has created new security and privacy challenges. Designing transparent, usable systems in support of personal privacy, security, and trust, requires advanced knowledge retrieval techniques that can support information sharing processes by applying appropriate policies. This paper proposes a Web 2.0 analysis methodology that provides reusable foundation components for information extraction, analysis and visualization. These components can be used by the same method can be also used by individuals to make a self-test of their Web 2.0 contributions and find out what inferences will be derived from their web presence.

Web 2.0 has changed the technological landscape of the Internet computing world today. The shift from traditional web which is also known as Web 1.0 is forced by the growing need for more efficient information sharing, collaboration, and business processes. The disclosure of personal/organizational information in Web 2.0 via social networks, digital contributions and data feeds has created new security and privacy challenges. Designing transparent, usable systems in support of personal privacy, security, and trust, requires advanced knowledge retrieval techniques that can support information sharing processes by applying appropriate policies. This paper proposes a Web 2.0 analysis methodology that provides reusable foundation components for information extraction, analysis and visualization. These components can be used by the same method can be also used by individuals to make a self-test of their Web 2.0 contributions and find out what inferences will be derived from their web presence.

We have developed a distributed DSS able to work in a dynamic way. This system is based on the use of mobile agents, which receive the user's queries and visit the DSS domains gathering the required information. One of the main elements to the efficient operation of this system is the router agent. Its responsibility is to give an adequate route to the mobile agents. In this paper we propose a structure for the router agent. It includes a Knowledge Base where the information managed by each domain is represented, and a mechanism of matching the query contents to the domain contents. We also present an interactive strategy to obtain the initial data to be stored in the Router Knowledge Base. The AHP method is used to support the task of evaluating the relative importance among terms representing each domain.

Traditional case-based reasoning uses a table/frame or scenario to represent a case. It assumed that similar input/event results in similar output/event state. However, similar cases may not have similar output/event states since problem solver may have different way to break down the problem. Thus, authors previously proposed problem-based case reasoning to overcome the limitation of the traditional approaches and used clustered ontology to represent the problem spaces of a case. However, synonym problem causes the mismatch of similar sub-problems of historical cases for new case. Thus, this paper proposed ontology-based similarity measurement to retrieve the similar sub-problems that overcomes the synonym problems on case retrieval. The recall and precise of ontology-based similarity measurement were higher than that of the traditional similarity measurement.

This comprehensive article delves into the emergence and potential of RAG (Retrieve, Augment, Generate) models as a significant advancement in Large Language Models (LLMs). The article examines how RAG models address key limitations of traditional LLMs by integrating dynamic knowledge retrieval mechanisms, enhancing relevance, creativity, and reliability in AI-generated content. It discusses the threecomponent architecture of RAG models, their ability to mitigate hallucinations, and their applications across various domains. The article also outlines future prospects, including multilingual capabilities, multimodal integration, and ethical considerations, positioning RAG models as a transformative force in natural language AI

Context: Software architecture documentation is used to communicate architectural knowledge. It is often difficult for document users to find all the architectural knowledge they need to do their tasks, and this results in wasted time and mistakes during development. Objective: In this paper we investigate how ontology-based documentation may support users in finding the architectural knowledge they need. Method: We executed a controlled experiment to test for differences in knowledge retrieval efficiency and effectiveness between two groups of master students that used two ontologies built from different understandings of the architectural knowledge needs of document users. Results: Use of the ontology built based on a better understanding of architectural knowledge needs was significantly more efficient or effective for retrieving part of the knowledge needed by document users. We analysed participants' search actions and identified which organisation of knowledge in the ontologies resulted in efficient and effective knowledge retrieval. Conclusion: We found that an improved understanding of knowledge needs allows for the construction of an ontology from which document users retrieve knowledge more efficiently and effectively. In some cases we found that the ontology support for knowledge needs had to be traded off against ontology design criteria.

A common approach to software architecture documentation in industry projects is the use of file-based documents. This approach offers a single-dimensional perspective on the architectural knowledge contained. Knowledge retrieval from file-based architecture documentation is efficient if the perspective chosen fits the needs of the readers; it is less so if the perspective does not match the needs of the readers. In this paper we describe an approach aimed at addressing architecture documentation retrieval issues. We have employed a software ontology in a semantic wiki optimized for architecture documentation. We have evaluated this ontologybased approach in a controlled industry experiment involving software professionals. The efficiency and effectiveness of the proposed approach is found to be better than that of the filebased approach.

The interaction of technology with humans have many adverse effects. The rapid growth and outreach of the social media and the Web have led to the dissemination of questionable and untrusted content among a wider audience, which has negatively influenced their lives and judgment. Different election campaigns around the world highlighted how "fake news"-misinformation that looks genuine-can be targeted towards specific communities to manipulate and confuse them. Ever since, automatic fake news detection has gained widespread attention from the scientific community. As a result, many research studies have been conducted to tackle the detection and spreading of fake news. While the first step of such tasks would be to classify claims associated based on their credibility, the next steps would involve identifying hidden patterns in style, syntax, and content of such news claims. We provide a comprehensive overview of what has already been done in this domain and other similar fields, and then propose a generalized method based on Deep Neural Networks to identify if a given claim is fake or genuine. By using different features like the authenticity of the source, perceived cognitive authority, style, and content-based factors, and natural language features, it is possible to predict fake news accurately. We have used a modular approach by combining techniques from information retrieval, natural language processing, and deep learning. Our classifier comprises two main sub-modules. The first sub-module uses the claim to retrieve relevant articles from the knowledge base which can then be used to verify the truth of the claim. It also uses word-level features for prediction. The second sub-module uses a deep neural network to learn the underlying style of fake content. Our experiments conducted on benchmark datasets show that for the given classification task we can obtain up to 82.4% accuracy by using a combination of two models; the first model was up to 72% accurate while the second model was around 81% accurate. Our detection model has the potential to automatically detect and prevent the spread of fake news, thus, limiting the caustic influence of technology in the human lives.

The interaction of technology with humans have many adverse effects. The rapid growth and outreach of the social media and the Web have led to the dissemination of questionable and untrusted content among a wider audience, which has negatively influenced their lives and judgment. Different election campaigns around the world highlighted how "fake news"-misinformation that looks genuine-can be targeted towards specific communities to manipulate and confuse them. Ever since, automatic fake news detection has gained widespread attention from the scientific community. As a result, many research studies have been conducted to tackle the detection and spreading of fake news. While the first step of such tasks would be to classify claims associated based on their credibility, the next steps would involve identifying hidden patterns in style, syntax, and content of such news claims. We provide a comprehensive overview of what has already been done in this domain and other similar fields, and then propose a generalized method based on Deep Neural Networks to identify if a given claim is fake or genuine. By using different features like the authenticity of the source, perceived cognitive authority, style, and content-based factors, and natural language features, it is possible to predict fake news accurately. We have used a modular approach by combining techniques from information retrieval, natural language processing, and deep learning. Our classifier comprises two main sub-modules. The first sub-module uses the claim to retrieve relevant articles from the knowledge base which can then be used to verify the truth of the claim. It also uses word-level features for prediction. The second sub-module uses a deep neural network to learn the underlying style of fake content. Our experiments conducted on benchmark datasets show that for the given classification task we can obtain up to 82.4% accuracy by using a combination of two models; the first model was up to 72% accurate while the second model was around 81% accurate. Our detection model has the potential to automatically detect and prevent the spread of fake news, thus, limiting the caustic influence of technology in the human lives.

There is an eternal problem for enterprise content management: "Everyday living is too fast, too busy, too complicated. More than at any time in history, it's important to have good information on just about every aspect of life. And, An onto-thesaurus based document and content management platform is very successful. It provides an improvement of recall and precision of information retrieval by ISO mono-lingual or multilingual thesauri indexing, homonymy resolution and onto-thesaurus.

This study investigates the relationship between knowledge management and employee effectiveness of deposit money banks in Nigeria using cross-sectional survey. Sixty branch managers from ten deposit money banks operating in Abakaliki, Ebonyi State were surveyed using convenience sampling technique. Forty four copies of questionnaire were filled Statistical tool used to analyse the hypotheses is Pearson Product Moment Correlation Coefficient (rs). Respondents’ profiles were analysed using descriptive statistics. The study found that knowledge management has positive significant association with employee effectiveness in deposit money banks in Abakaliki. It concluded that knowledge management measured in terms of knowledge acquisition, knowledge dissemination, knowledge conservation and knowledge retrieval enhances employee effectiveness when resources are provided for the employees quite apart from setting goals on how jobs should be carried out.  Based on the conclusion above, one ...

Background and aim: The combination of ontology-based retrieval systems leads to the semantic retrieval of words. The aim of this study was to review ontology articles in information retrieval using scientometric techniques. Materials and methods: The present study was conducted using the documentary method and word cluster analysis. The research population comprised 2595 articles indexed in two databases, Scopus and Web of Science, from 2001 to 2023. The data were analyzed using Excel, BibExcel, SPSS 26 and UCINET. VOSviewer was used to draw research maps. Findings: The growth of articles in ontology and information retrieval was low and the annual growth rate averaged 0.11%.Stanford and California universities were the most prolific organizations, contributing to 6 articles, and the field of computer science was the most prolific with 43% of the articles written. The word clustering led to the formation of 4 thematic clusters: semantic retrieval of information, non-human ontology, classification of systems, and role of technology. In addition, there was a positive correlation between science production and centralities (degree centrality 0.323, closeness centrality 0.278, and betweenness centrality 0.447). Conclusion: The evolution of the words used in the articles has shown that although the growth of article production in this field has increased from the beginning, the development of ontology technologies in information retrieval started with a weak semantic system called information classification, and after the various stages of development, it now uses machine learning to understand user requirements and process information with the help of artificial intelligence.

Badges are endemic to online interaction sites, from Question and Answer (Q&A) websites to ride sharing, as systems for rewarding participants for their contributions. This paper studies how badge design affects people's contributions and behavior over time. Past work has shown that badges "steer" people's behavior toward substantially increasing the amount of contributions before obtaining the badge, and immediately decreasing their contributions thereafter, returning to their baseline contribution levels. In contrast, we find that the steering effect depends on the type of user, as modeled by the rate and intensity of the user's contributions. We use these measures to distinguish between different groups of user activity, including users who are not affected by the badge system despite being significant contributors to the site. We provide a predictive model of how users change their activity group over the course of their lifetime in the system. We demonstrate our approach empirically in three different Q&A sites on Stack Exchange with hundreds of thousands of users, and we discuss the implications for system designers.

The amount of multimedia content produced and made available on the World Wide Web, and in professional and, not least, personal collections, is constantly growing, resulting in equally increasing needs in terms of efficient and effective ways to access it. Enabling smooth access at a level that meets user expectations and needs has been the holy grail in content-based retrieval for decades as it is intertwined with the so-called semantic gap between the features that can be extracted from such content through automatic analysis and the conveyed meaning as perceived by the end users. Numerous efforts towards more reliable and effective visual content analysis that target the extraction of user-oriented content descriptions have been reported, addressing a variety of domains and applications, and following diverse methodologies. Among the reported literature, knowledge-based approaches utilising explicit, a priori, knowledge constitute a popular choice aiming at analysis methods decoupled from application-specific implementations. Such knowledge may address various aspects including visual characteristics and numerical representations, topological knowledge about the examined domain, contextual knowledge, as well as knowledge driving the selection and execution of the processing steps required. Among the different knowledge representations adopted in the reported literature, ontologies, being the key enabling technology of the Semantic Web (SW) vision for knowledge sharing and reuse through machine processable metadata, have been favoured in recent efforts. Indicative state-of-the-art approaches include, among others, the work presented in Little and Hunter (2004), and Hollink, Little and Hunter (2005), where ontologies have been used to represent objects of the

Knowledge is increasingly recognised as an important asset in the construction industry. It has been critically important as the drivers for success for a knowledge driven business such as quantity surveying (QS) practices. However, the recent turnover trends among the QS employee is quite alarming and that the ‘culture’ will lead to loss to the QS firm in terms of manpower investments and loss of tacit knowledge. Thus, this research was carried out to determine the perception of the QS practices’ employers on the importance of their knowledgeable employees and also to identify the strategies adopted by them in retaining their employees’ knowledge. Semi structured interview was conducted among the QS firms’ employers and their answers were recorded and transcribed. The findings revealed that the employers of the QS practices do regard their employees’ knowledge as very important to them. In return, to retain their knowledgeable employees, several approaches have been taken namely in...

Purpose: The purpose of this paper is to present different issues facing the problem of knowledge retention by construction sector companies in the UAE. Methodology: Existing framework in the area of knowledge retention has been used to assess three large construction consultancies in the UAE. The case study methodology used in this paper highlights some key issues in the area of knowledge retention in the UAE. Findings: Based on the analysis of the knowledge retention system the major drivers for its successful implementation are prevalence of a culture of sharing knowledge, reward and recognition for sharing knowledge, a technology platform that can accommodate multi formats of files, awareness of knowledge retention system and its benefits among its employees, and top management support. Research Limitations: The paper presents three case studies and draws conclusions about the UAE construction sector. Although the three companies are large companies, there are several Small and ...

There is an ever increasing growth in the use of Q&A websites such as Stack Overflow (SO), so are the number of posts on them. These websites serve as knowledge sharing platforms where Subject Matter Experts (SMEs) and developers answer questions posted by other users. It is effort intensive for developers to navigate to right posts because of the large volume of posts on the platform, despite the presence of existing tags, that are based on technologies. Tagging these posts based on their context and purpose might help developers and SMEs in easily identifying questions they wish to answer and also in identifying contextually similar posts. To support this idea, we propose SOTagger as a prototype plug-in for Stack Overflow to tag questions contextually. We have considered SO data provided on SOTorrent and automated the identification of 6 categories of questions using Latent Dirichlet Allocation. We have also manually verified relevance of these categories. Using these categories and dataset, we have built a classification model to classify a post into one of these six categories using Support Vector Machine. We have evaluated SOTagger by conducting a user survey with 32 developers. The preliminary results are promising with about 80% developers recommending the plugin to others.

The International Conference on Information Retrieval and Knowledge Management (CIKM) brings together three avenues of data-oriented research, namely, Database Management, Information Retrieval and Knowledge Management. The confluence of these avenues becomes evident also in the PhD theses of doctoral students: Stream processing makes use of knowledge representation techniques, linked data is emerging as a research topic that bridges information retrieval and knowledge representation, and new forms of querying draw on techniques from both information retrieval and databases. In this paper, we survey new PhD theses at the meeting point of the three research avenues. Our survey is based on the 5th PhD workshop at the ACM CIKM conference. The topics include themes as diverse as link prediction, source code querying, and video stream processing.

Knowledge is increasingly recognised as an important asset in the construction industry. It has been critically important as the drivers for success for a knowledge driven business such as quantity surveying (QS) practices. However, the recent turnover trends among the QS employee is quite alarming and that the ‘culture’ will lead to loss to the QS firm in terms of manpower investments and loss of tacit knowledge. Thus, this research was carried out to determine the perception of the QS practices’ employers on the importance of their knowledgeable employees and also to identify the strategies adopted by them in retaining their employees’ knowledge. Semi structured interview was conducted among the QS firms’ employers and their answers were recorded and transcribed. The findings revealed that the employers of the QS practices do regard their employees’ knowledge as very important to them. In return, to retain their knowledgeable employees, several approaches have been taken namely in...

With employees leaving construction organizations for all sorts of reasons, it is imperative that organizations look at ways to retain their knowledge. This paper presents a four step process in order to ensure knowledge retention. The four steps include: Socialization, codification, ...

The use of the Internet around the world has skyrocketed in recent years, and online forums have developed to accommodate interests, both broad and specialized, as people have begun to seek advice and knowledge from the vast resource of Internet users rather than counting on the relatively small number of people they personally know. Question-and-answer (Q&A) platforms such as Yahoo! Answers and Quora allow users to ask questions regarding any topic, while Stack Overflow is dedicated to the specific field of computer science.

This paper presents a new type of fuzzy sets, called "Hierarchical Fuzzy Sets", that apply when the considered domain of values is not "flat", but contains values that are more specific than others according to the "kind of" relation. We study the properties of such fuzzy sets, that can be defined in a short way on a part of the hierarchy, or exhaustively (by their "closure") on the whole hierarchy. We show that hierarchical fuzzy sets form equivalence classes in regard to their closures and that each class has a particular representative called "minimal fuzzy set". We propose a use of this minimal fuzzy set for query enlargement purposes and thus present a methodology for hierarchical fuzzy set generalization. We finally present an experimental evaluation of the theoretical results described in the paper, in a practical application.

Retrieving accurate information from the Web is a great challenge to users. The existing information retrieval systems are mostly term-based and thus need to be enhanced toward knowledge-based. User information needs need to be better captured in order to deliver personalized search results. In this paper, an ontology-based framework is proposed for capturing user information needs using a world knowledge base and the user's local instance repository. The framework aims to discover a user's background knowledge for knowledge retrieval. The evaluation result is encouraging, in which the proposed model achieved the same performance as a manual user model.

Technical Q&A sites have become essential for software engineers as they constantly seek help from other experts to solve their work problems. Despite their success, many questions remain unresolved, sometimes because the asker does not acknowledge any helpful answer. In these cases, an information seeker can only browse all the answers within a question thread to assess their quality as potential solutions. We approach this time-consuming problem as a binary-classification task where a best-answer prediction model is built to identify the accepted answer among those within a resolved question thread, and the candidate solutions to those questions that have received answers but are still unresolved. In this paper, we report on a study aimed at assessing 26 best-answer prediction models in two steps. First, we study how models perform when predicting best answers in Stack Overflow, the most popular Q&A site for software engineers. Then, we assess performance in a cross-platform setting where the prediction models are trained on Stack Overflow and tested on other technical Q&A sites. Our findings show that the choice of the classifier and automatied parameter tuning have a large impact on the prediction of the best answer. We also demonstrate that our approach to the bestanswer prediction problem is generalizable across technical Q&A sites. Finally, we provide practical recommendations to Q&A platform designers to curate and preserve the crowdsourced knowledge shared through these sites.

Computer scientists and engineers work with increasing amounts of information. These data are used for knowledge retrieval, data management decision support and so on. Sorting algorithms are important procedures that if efficiently composed and implemented can increase speed of data processing and decision correctness. Many various sorting algorithms and their modifications are applicable in large computer systems. However, as the computer architectures are more efficient with each release and the software is more complex with each version there is need to improve sorting methods applied for big data computation.

We have developed a distributed DSS capable to working in a dynamic way. That is, when a domain of an organization needs a new kind of information, the system looks for this information. This system is based on the usage of mobile agents, which receive the user's queries and visit the appropriate DSS domains to gather the required information. The system itself must analyze where this information can be generated. To make this decision there is an intelligent agent (the Router) with a knowledge base (KB) where the information managed by each domain is represented. In this work, we present a strategy to obtain the initial data to be stored in the KB, a knowledge retrieval mechanism from the KB, and a learning mechanism so that the KB and the DSS operation can be continually improved. The proposed learning process is an interpretative case-based reasoning, which uses a set of rules to analyze the results of the information retrieval process and modifies the content of the router K...

We studied the problem of searching answers for questions on a Question-and-Answer Website from knowledge bases. A number of research e↵orts had been developed using Stack Overflow data, which is available for the public. Surprisingly, only a few papers tried to improve the search for better answers. Furthermore, current approaches for searching a Question-and-Answer Website are usually limited to the question database, which is usually the website own content. We showed it is feasible to use knowledge bases as sources for answers. We implemented both vector-space and topic-space representations for our datasets and compared these distinct techniques. Finally, we proposed a hybrid ranking approach that took advantage of a machine-learned classifier to incorporate the tag information into the ranking and showed that it was able to improve the retrieval performance.

We studied the problem of searching answers for questions on a Question-and-Answer Website from knowledge bases. A number of research e↵orts had been developed using Stack Overflow data, which is available for the public. Surprisingly, only a few papers tried to improve the search for better answers. Furthermore, current approaches for searching a Question-and-Answer Website are usually limited to the question database, which is usually the website own content. We showed it is feasible to use knowledge bases as sources for answers. We implemented both vector-space and topic-space representations for our datasets and compared these distinct techniques. Finally, we proposed a hybrid ranking approach that took advantage of a machine-learned classifier to incorporate the tag information into the ranking and showed that it was able to improve the retrieval performance.

Purpose: The purpose of this paper is to present different issues facing the problem of knowledge retention by construction sector companies in the UAE. Methodology: Existing framework in the area of knowledge retention has been used to assess three large construction consultancies in the UAE. The case study methodology used in this paper highlights some key issues in the area of knowledge retention in the UAE. Findings: Based on the analysis of the knowledge retention system the major drivers for its successful implementation are prevalence of a culture of sharing knowledge, reward and recognition for sharing knowledge, a technology platform that can accommodate multi formats of files, awareness of knowledge retention system and its benefits among its employees, and top management support. Research Limitations: The paper presents three case studies and draws conclusions about the UAE construction sector. Although the three companies are large companies, there are several Small and ...

A common approach to software architecture documentation in industry projects is the use of file-based documents. This approach offers a single-dimensional arrangement of the architectural knowledge. Knowledge retrieval from file-based architecture documentation is efficient if the organisation of knowledge supports the needs of the readers; otherwise it can be difficult. In this paper, we compare the organisation and retrieval of architectural knowledge in a file-based documentation approach and an ontology-based documentation approach. The ontology-based approach offers a multi-dimensional organisation of architectural knowledge by means of a software ontology and semantic wiki, whereas file-based documentation typically uses hierarchical organisation by directory structure and table of content. We conducted case studies in two companies to study the efficiency and effectiveness of retrieving architectural knowledge from the different organisations of knowledge. We found that the use of better knowledge organisation correlates with the efficiency and effectiveness of AK retrieval. Professionals who used the knowledge organisation found this beneficial.

Context: Software architecture documentation is used to communicate architectural knowledge. It is often difficult for document users to find all the architectural knowledge they need to do their tasks, and this results in wasted time and mistakes during development. Objective: In this paper we investigate how ontology-based documentation may support users in finding the architectural knowledge they need. Method: We executed a controlled experiment to test for differences in knowledge retrieval efficiency and effectiveness between two groups of master students that used two ontologies built from different understandings of the architectural knowledge needs of document users. Results: Use of the ontology built based on a better understanding of architectural knowledge needs was significantly more efficient or effective for retrieving part of the knowledge needed by document users. We analysed participants' search actions and identified which organisation of knowledge in the ontologies resulted in efficient and effective knowledge retrieval. Conclusion: We found that an improved understanding of knowledge needs allows for the construction of an ontology from which document users retrieve knowledge more efficiently and effectively. In some cases we found that the ontology support for knowledge needs had to be traded off against ontology design criteria.