Academic literature on the topic 'Semantic search algorithms'

The advancement of semantic search algorithms relies heavily on the integration of sophisticated mathematical frameworks to decipher and interpret the semantics of user queries and web documents. This article provides an in-depth exploration of three key mathematical models utilized in semantic search: Vector Space Models (VSM), Latent Semantic Analysis (LSA), and Word Embeddings. Each model is meticulously examined, elucidating their mathematical foundations, operational principles, and integration into semantic search algorit hms. From the mathematical representation of documents and queries in vector space to the application of Singular Value Decomposition (SVD) in uncovering latent semantic structures, the article delves into the intricacies of these models. Furthermore, it explores how Word Embeddings, exemplified by Word2Vec and GloVe, revolutionize semantic understanding through dense vector representations of words. By synthesizing these mathematical frameworks into semantic search algorithms, search engines can bridge the semantic gap between user intent and search results, ultimately enhancing the accuracy, relevance, and user experience of information retrieval. Through this nuanced analysis, the article underscores the indispensable role of mathematics in propelling the evolution of semantic search technology towards more intuitive and efficient information retrieval systems in the digital

Code search uses natural language queries to retrieve code snippets from a vast database, identifying those that are semantically similar to the query. This enables developers to reuse code and enhance software development efficiency. Most existing code search algorithms focus on capturing semantic and structural features by learning from both text and code graph structures. However, these algorithms often struggle to capture deeper semantic and structural features within these sources, leading to lower accuracy in code search results. To address this issue, this paper proposes a novel semantics-enhanced neural code search algorithm called SENCS, which employs graph serialization and a two-stage attention mechanism. First, the code program dependency graph is transformed into a unique serialized encoding, and a bidirectional long short-term memory (LSTM) model is used to learn the structural information of the code in the graph sequence to generate code vectors rich in structural features. Second, a two-stage attention mechanism enhances the embedded vectors by assigning different weight information to various code features during the code feature fusion phase, capturing significant feature information from different code feature sequences, resulting in code vectors rich in semantic and structural information. To validate the performance of the proposed code search algorithm, extensive experiments were conducted on two widely used code search datasets, CodeSearchNet and JavaNet. The experimental results show that the proposed SENCS algorithm improves the average code search accuracy metrics by 8.30 % (MRR) and 17.85% (DCG) and compared to the best baseline code search model in the literature, with an average improvement of 14.86% in the SR@1 metric. Experiments with two open-source datasets demonstrate SENCS achieves a better search effect than state of-the-art models.

Search for documents is a common and pertinent task lots of organizations face every day as well as common Internet users in their daily searches. One specific document search is scientific paper search in reference manager systems such as Mendeley or IEEExplore. Considering the difficult task finding documents can sometimes represent, semantic search is currently being applied to improve this type of search. As the act of deciding if a document is a good result for a given search expression is vague, fuzziness becomes an important aspect when defining search algorithms. In this paper, the author present a fuzzy algorithm for improving documental searches optimized for specific scenarios where we want to find a document but don´t remember the exact words used, if plural or singular words were used or if a synonym was used. The author also present the application of this algorithm to a real scenario comparing to Mendeley and IEEExplore results.

With the Development of web service technology, a single web service cannot fulfill different users’ diverse requirements. Adding semantic information to the input-output message of web services provides us a method to implement web service composition automatically. After researching on existing algorithms for web service composition, this article proposed a QoS-oriented web service composition algorithm based on graph search with semantic information.

Clustering is an important unsupervised analysis technique for big data mining. It finds its application in several domains including biomedical documents of the MEDLINE database. Document clustering algorithms based on metaheuristics is an active research area. However, these algorithms suffer from the problems of getting trapped in local optima, need many parameters to adjust, and the documents should be indexed by a high dimensionality matrix using the traditional vector space model. In order to overcome these limitations, in this paper a new documents clustering algorithm (ASOS-LSI) with no parameters is proposed. It is based on the recent symbiotic organisms search metaheuristic (SOS) and enhanced by an acceleration technique. Furthermore, the documents are represented by semantic indexing based on the famous latent semantic indexing (LSI). Conducted experiments on well-known biomedical documents datasets show the significant superiority of ASOS-LSI over five famous algorithms in terms of compactness, f-measure, purity, misclassified documents, entropy, and runtime.

Users often struggle to discover the information they need online because of the massive volume of data that is readily available as well as being generated every day in the today’s digital age. Traditional keyword-based search engines may not be able to handle complex queries, which could result in irrelevant or insufficient search results. This issue can be solved by semantic search, which utilises machine learning and natural language processing to interpret the meaning and context of a user's query. In this paper we focus on analyzing the BM-25 algorithm, Mean of Word Vectors approach, Universal Sentence Encoder model, and Sentence-BERT model on the CISI Dataset for Semantic Search Task. The results indicate that, the Finetuned SBERT model performs the best.

The emergence of multiple web pages day by day leads to the development of the semantic web technology. A World Wide Web Consortium (W3C) standard for storing semantic web data is the resource description framework (RDF). To enhance the efficiency in the execution time for querying large RDF graphs, the evolving metaheuristic algorithms become an alternate to the traditional query optimization methods. This paper focuses on the problem of query optimization of semantic web data. An efficient algorithm called adaptive Cuckoo search (ACS) for querying and generating optimal query plan for large RDF graphs is designed in this research. Experiments were conducted on different datasets with varying number of predicates. The experimental results have exposed that the proposed approach has provided significant results in terms of query execution time. The extent to which the algorithm is efficient is tested and the results are documented.

Given a set of documents and an input query that is expressed in a natural language, the problem of document search is retrieving the most relevant documents. Unlike most existing systems that perform document search based on keyword matching, we propose a method that considers the meaning of the words in the queries and documents. As a result, our algorithm can return documents that have no words in common with the input query as long as the documents are relevant. For example, a document that contains the words "Ford", "Chrysler" and "General Motors" multiple times is surely relevant for the query "car" even if the word "car" never appears in the document. Our information retrieval algorithm is based on a similarity graph that contains the degree of semantic closeness between terms, where a term can be a word or a phrase. Since the algorithms that constructs the similarity graph takes as input a myriad of parameters, in this paper we fine-tune the part of the algorithm that constructs the Wikipedia part of the graph. Specifically, we experimentally fine-tune the algorithm on the Miller and Charles study benchmark that contains 30 pairs of terms and their similarity score as determined by human users. We then evaluate the performance of the fine-tuned algorithm on the Cranfield benchmark that contains 1400 documents and 225 natural language queries. The benchmark also contains the relevant documents for every query as determined by human judgment. The results show that the fine-tuned algorithm produces higher mean average precision (MAP) score than traditional keyword-based search algorithms because our algorithm considers not only the words and phrases in the query and documents, but also their meaning.

The rapid developing international of internet, Semantic Web has become a platform for intelligent agents mainly in the healthcare sector. Inside the beyond few years there is a widening in the Semantic web data field in the healthcare industry. With a growth in the quantity of Semantic web data field in health industry, there exist some challenges to be resolved. One such challenge is to provide an efficient querying mechanism that can handle large number of Semantic web data. Consider many query languages; especially SPARQL (SPARQL Protocol and RDF Query Language) is the most popular query language. Each of these query languages has their own design strategy and it was identified in research that it is difficult to handle and query large quantity of RDF data efficiently using these languages. In the proposed process, Harmony search identify met heuristic algorithm to optimize the SPARQL queries in the healthcare data in the applicable manner. The application of Harmony search algorithm is evaluated with large Resource Description Framework (RDF) datasets and SPARQL queries. To assess performance, the algorithm’s implementation is compared to existing nature-inspired algorithms. The performance analysis shows that the proposed application performs well for large RDF datasets.

Universal Domain Adaptation (UniDA) focuses on transferring source domain knowledge to the target domain under both domain shift and unknown category shift. Its main challenge lies in identifying common class samples and aligning them. Current methods typically obtain target domain semantics centers from an unconstrained continuous image representation space. Due to domain shift and the unknown number of clusters, these centers often result in complex and less robust alignment algorithm. In this paper, based on vision-language models, we search for semantic centers in a semantically meaningful and discrete text representation space. The constrained space ensures almost no domain bias and appropriate semantic granularity for these centers, enabling a simple and robust adaptation algorithm. Specifically, we propose TArget Semantics Clustering (TASC) via Text Representations, which leverages information maximization as a unified objective and involves two stages. First, with the frozen encoders, a greedy search-based framework is used to search for an optimal set of text embeddings to represent target semantics. Second, with the search results fixed, encoders are refined based on gradient descent, simultaneously achieving robust domain alignment and private class clustering. Additionally, we propose Universal Maximum Similarity (UniMS), a scoring function tailored for detecting open-set samples in UniDA. Experimentally, we evaluate the universality of UniDA algorithms under four category shift scenarios. Extensive experiments on four benchmarks demonstrate the effectiveness and robustness of our method, which has achieved state-of-the-art performance.

Searching is a fundamental task to support research. Current search engines are keyword-based. Semantic technologies promise a next generation of semantic search engines, which will be able to answer questions. Current approaches either apply natural language processing to unstructured text or they assume the existence of structured statements over which they can reason. This work provides a system for combining the classical keyword-based search engines with semantic annotation. Conventional search results are annotated using a customized annotation algorithm, which takes the textual properties and requirements such as speed and scalability into account. The biomedical background knowledge consists of the GeneOntology and Medical Subject Headings and other related entities, e.g. proteins/gene names and person names. Together they provide the relevant semantic context for a search engine for the life sciences. We develop the system GoWeb for semantic web search and evaluate it using three benchmarks. It is shown that GoWeb is able to aid question answering with success rates up to 79%. Furthermore, the system also includes semantic hyperlinks that enable semantic browsing of the knowledge space. The semantic hyperlinks facilitate the use of the eScience infrastructure, even complex workflows of composed web services. To complement the web search of GoWeb, other data source and more specialized information needs are tested in different prototypes. This includes patents and intranet search. Semantic search is applicable for these usage scenarios, but the developed systems also show limits of the semantic approach. That is the size, applicability and completeness of the integrated ontologies, as well as technical issues of text-extraction and meta-data information gathering. Additionally, semantic indexing as an alternative approach to implement semantic search is implemented and evaluated with a question answering benchmark. A semantic index can help to answer questions and address some limitations of GoWeb. Still the maintenance and optimization of such an index is a challenge, whereas GoWeb provides a straightforward system.

Représenter l'information décrivant la musique est une activité complexe, qui implique différentes sous-tâches. Ce manuscrit de thèse porte principalement sur la musique classique et étudie comment représenter et exploiter ses informations. L'objectif principal est l'étude de stratégies de représentation et de découverte des connaissances appliquées à la musique classique, dans des domaines tels que la production de base de connaissances, la prédiction de métadonnées et les systèmes de recommandation. Nous proposons une architecture pour la gestion des métadonnées de musique à l'aide des technologies du Web Sémantique. Nous introduisons une ontologie spécialisée et un ensemble de vocabulaires contrôlés pour les différents concepts spécifiques à la musique. Ensuite, nous présentons une approche de conversion des données, afin d’aller au-delà de la pratique bibliothécaire actuellement utilisée, en s’appuyant sur des règles de mapping et sur l’interconnexion avec des vocabulaires contrôlés. Enfin, nous montrons comment ces données peuvent être exploitées. En particulier, nous étudions des approches basées sur des plongements calculés sur des métadonnées structurées, des titres et de la musique symbolique pour classer et recommander de la musique. Plusieurs applications de démonstration ont été réalisées pour tester les approches et les ressources précédentes<br>Representing the information about music is a complex activity that involves different sub-tasks. This thesis manuscript mostly focuses on classical music, researching how to represent and exploit its information. The main goal is the investigation of strategies of knowledge representation and discovery applied to classical music, involving subjects such as Knowledge-Base population, metadata prediction, and recommender systems. We propose a complete workflow for the management of music metadata using Semantic Web technologies. We introduce a specialised ontology and a set of controlled vocabularies for the different concepts specific to music. Then, we present an approach for converting data, in order to go beyond the librarian practice currently in use, relying on mapping rules and interlinking with controlled vocabularies. Finally, we show how these data can be exploited. In particular, we study approaches based on embeddings computed on structured metadata, titles, and symbolic music for ranking and recommending music. Several demo applications have been realised for testing the previous approaches and resources

L'utilisation des téléphones intelligents, le recours aux assistants personnels intelligents ou encore le développement des maisons intelligentes sont autant d'exemples illustrant le développement toujours plus rapide de l'informatique ubiquitaire, de l'Internet des objets et de l'intelligence artificielle. Le croisement des résultats issus de ces domaines de recherche contribue à changer notre quotidien et constitue un environnement fertile pour de nouveaux travaux. Ainsi, l’intégration des entités cyber-physiques dans des environnements sociaux de différentes échelles allant des maisons aux villes intelligentes amène de très nombreuses perspectives. Ce changement de paradigme met à notre disposition une énorme quantité d'informations et de services utiles, offrant ainsi la possibilité de traiter les problèmes classiques de manière nouvelle, différente et potentiellement plus efficace. Si les solutions à construire bénéficient de ces possibilités, elles doivent également répondre à de nouvelles contraintes et nouveaux défis. La recherche d’itinéraires multi-objectifs est un sous-cas du problème classique de recherche d'un chemin entre un lieu de départ et une destination auquel s'ajoute la contrainte de passage par un ensemble de lieux permettant de satisfaire un ensemble de buts. L'objectif de cette thèse est de proposer une solution pour la résolution de la recherche d'itinéraires multi-objectifs appliqués aux environnements cyber-physiques tels que les Smart Transits. Dans notre solution, nous avons proposé une méthode fondée sur les technologies du web sémantique pour modéliser de manière intégrée un environnement cyber-physique dans toutes ses dimensions, i.e., cybernétiques, physiques et sociales. Pour la recherche de chemin, nous avons proposé une approche multi-agents, exécutant un algorithme de recherche collaborative et incrémentale, qui utilise les connaissances de l'environnement pour trouver le chemin optimal. Cet algorithme adapte aussi le chemin en prenant en compte la dynamique de l'environnement<br>From intelligent artificial personal assistants to smart cities, we are experiencing the shifting towards Internet of Things (IoT), ubiquitous computing, and artificial intelligence. Cyber-physical entities are embedded in social environments of various scales from smart homes, to smart airports, to smart cities, and the list continues.This paradigm shift coupled with ceaseless expansion of the Web supplies us with tremendous amount of useful information and services, which creates opportunities for classical problems to be addressed in new, different, and potentially more efficient manners. Along with the new possibilities, we are, at the same time, presented with new constraints, problems, and challenges. Multi-goal pathfinding, a variant of the classical pathfinding, is a problem of finding a path between a start and a destination which also allows a set of goals to be satisfied along the path. The aim of this dissertation is to propose a solution to solve multi-goal pathfinding in ubiquitous environments such as smart transits. In our solution, to provide an abstraction of the environment, we proposed a knowledge model based on the semantic web technologies to describe a ubiquitous environment integrating its cybernetic, physical, and social dimensions. To perform the search, we developed a multi-agent algorithm based on a collaborative and incremental search algorithm that exploits the knowledge of the environment to find the optimal path. The proposed algorithm continuously adapts the path to take into account the dynamics of the environment

<p>For researches in the field of the data mining and machine learning the necessary condition is an availability of various input data set. Now researchers create the databases of such sets. Examples of the following systems are: The UCI Machine Learning Repository, Data Envelopment Analysis Dataset Repository, XMLData Repository, Frequent Itemset Mining Dataset Repository. Along with above specified statistical repositories, the whole pleiad from simple filestores to specialized repositories can be used by researchers during solution of applied tasks, researches of own algorithms and scientific problems. It would seem, a single complexity for the user will be search and direct understanding of structure of so separated storages of the information. However detailed research of such repositories leads us to comprehension of deeper problems existing in usage of data. In particular a complete mismatch and rigidity of data files structure with SDMX - Statistical Data and Metadata Exchange - standard and structure used by many European organizations, impossibility of preliminary data origination to the concrete applied task, lack of data usage history for those or other scientific and applied tasks.</p><p>Now there are lots of methods of data miming, as well as quantities of data stored in various repositories. In repositories there are no methods of DM (data miming) and moreover, methods are not linked to application areas. An essential problem is subject domain link (problem domain), methods of DM and datasets for an appropriate method. Therefore in this work we consider the building problem of ontological models of DM methods, interaction description of methods of data corresponding to them from repositories and intelligent agents allowing the statistical repository user to choose the appropriate method and data corresponding to the solved task. In this work the system structure is offered, the intelligent search agent on ontological model of DM methods considering the personal inquiries of the user is realized.</p><p>For implementation of an intelligent data and metadata exchange repository the agent oriented approach has been selected. The model uses the service oriented architecture. Here is used the cross platform programming language Java, multi-agent platform Jadex, database server Oracle Spatial 10g, and also the development environment for ontological models - Protégé Version 3.4.</p>

[EN] The long-lasting trend in the field of computation of stress and resource distribution has found its way into computer networks via the concept of peer-to-peer (P2P) connectivity. P2P is a symmetrical model, where each network node is enabled a comparable range of capacities and resources. It stands in a stark contrast to the classical, strongly asymmetrical client-server approach. P2P, originally considered only a complimentary, server-side structure to the straightforward client-server model, has been shown to have the substantial potential on its own, with multiple, widely known benefits: good fault tolerance and recovery, satisfactory scalability and intrinsic load distribution. However, contrary to client-server, P2P networks require sophisticated solutions on all levels, ranging from network organization, to resource location and managing. In this thesis we address one of the key issues of P2P networks: performing efficient resource searches of semantic nature under realistic, dynamic conditions. There have been numerous solutions to this matter, with evolutionary, stigmergy-based, and simple computational foci, but few attempt to resolve the full range of challenges this problem entails. To name a few: real-life P2P networks are rarely static, nodes disconnect, reconnect and change their content. In addition, a trivial incorporation of semantic searches into well-known algorithms causes significant decrease in search efficiency. In our research we build a solution incrementally, starting with the classic Ant Colony System (ACS) within the Ant Colony Optimization metaheuristic (ACO). ACO is an algorithmic framework used for solving combinatorial optimization problems that fits contractually the problem very well, albeit not providing an immediate solution to any of the aforementioned problems. First, we propose an efficient ACS variant in structured (hypercube structured) P2P networks, by enabling a path-post processing algorithm, which called Tabu Route Optimization (TRO). Next, we proceed to resolve the issue of network dynamism with an ACO-compatible information diffusion approach. Consequently, we attempt to incorporate the semantic component of the searches. This initial approximation to the problem was achieved by allowing ACS to differentiate between search types with the pheromone-per-concept idea. We called the outcome of this merger Routing Concept ACS (RC-ACS). RC-ACS is a robust, static multipheromone implementation of ACS. However, we were able to conclude from it that the pheromone-per-concept approach offers only limited scalability and cannot be considered a global solution. Thus, further progress was made in this respect when we introduced to RC-ACS our novel idea: dynamic pheromone creation, which replaces the static one-to-one assignment. We called the resulting algorithm Angry Ant Framework (AAF). In AAF new pheromone levels are created as needed and during the search, rather than prior to it. The final step was to enable AAF, not only to create pheromone levels, but to reassign them to optimize the pheromone usage. The resulting algorithm is called EntropicAAF and it has been evaluated as one of the top-performing algorithms for P2P semantic searches under all conditions.<br>[ES] La popular tendencia de distribución de carga y recursos en el ámbito de la computación se ha transmitido a las redes computacionales a través del concepto de la conectividad peer-to-peer (P2P). P2P es un modelo simétrico, en el cual a cada nodo de la red se le otorga un rango comparable de capacidades y recursos. Se trata de un fuerte contraste con el clásico y fuertemente asimétrico enfoque cliente-servidor. P2P, originalmente considerado solo como una estructura del lado del servidor complementaria al sencillo modelo cliente-servidor, ha demostrado tener un potencial considerable por sí mismo, con múltiples beneficios ampliamente conocidos: buena tolerancia a fallos y recuperación, escalabilidad satisfactoria y distribución de carga intrínseca. Sin embargo, al contrario que el modelo cliente-servidor, las redes P2P requieren de soluciones sofisticadas a todos los niveles, desde la organización de la red hasta la gestión y localización de recursos. Esta tesis aborda uno de los problemas principales de las redes P2P: la búsqueda eficiente de recursos de naturaleza semántica bajo condiciones dinámicas y realistas. Ha habido numerosas soluciones a este problema basadas en enfoques evolucionarios, estigmérgicos y simples, pero pocas han tratado de resolver el abanico completo de desafíos. En primer lugar, las redes P2P reales son raramente estáticas: los nodos se desconectan, reconectan y cambian de contenido. Además, la incorporación trivial de búsquedas semánticas en algoritmos conocidos causa un decremento significativo de la eficiencia de la búsqueda. En esta investigación se ha construido una solución de manera incremental, comenzando por el clásico Ant Colony System (ACS) basado en la metaheurística de Ant Colony Optimization (ACO). ACO es un framework algorítmico usado para búsquedas en grafos que encaja perfectamente con las condiciones del problema, aunque no provee una solución inmediata a las cuestiones mencionadas anteriormente. En primer lugar, se propone una variante eficiente de ACS para redes P2P estructuradas (con estructura de hipercubo) permitiendo el postprocesamiento de las rutas, al que hemos denominado Tabu Route Optimization (TRO). A continuación, se ha tratado de resolver el problema del dinamismo de la red mediante la difusión de la información a través de una estrategia compatible con ACO. En consecuencia, se ha tratado de incorporar el componente semántico de las búsquedas. Esta aproximación inicial al problema ha sido lograda permitiendo al ACS diferenciar entre tipos de búsquedas através de la idea de pheromone-per-concept. El resultado de esta fusión se ha denominado Routing Concept ACS (RC-ACS). RC-ACS es una implementación multiferomona estática y robusta de ACS. Sin embargo, a partir de esta implementación se ha podido concluir que el enfoque pheromone-per-concept ofrece solo escalabilidad limitada y que no puede ser considerado una solución global. Por lo tanto, para lograr una mejora a este respecto, se ha introducido al RC-ACS una novedosa idea: la creación dinámica de feromonas, que reemplaza la asignación estática uno a uno. En el algoritmo resultante, al que hemos denominado Angry Ant Framework (AAF), los nuevos niveles de feromona se crean conforme se necesitan y durante la búsqueda, en lugar de crearse antes de la misma. La mejora final se ha obtenido al permitir al AAF no solo crear niveles de feromona, sino también reasignarlos para optimizar el uso de la misma. El algoritmo resultante se denomina EntropicAAF y ha sido evaluado como uno de los algoritmos más exitosos para las búsquedas semánticas P2P bajo todas las condiciones.<br>[CAT] La popular tendència de distribuir càrrega i recursos en el camp de la computació s'ha estès cap a les xarxes d'ordinadors a través del concepte de connexions d'igual a igual (de l'anglès, peer to peer o P2P). P2P és un model simètric on cada node de la xarxa disposa del mateix nombre de capacitats i recursos. P2P, considerat originàriament només una estructura situada al servidor complementària al model client-servidor simple, ha provat tindre el suficient potencial per ella mateixa, amb múltiples beneficis ben coneguts: una bona tolerància a errades i recuperació, una satisfactòria escalabilitat i una intrínseca distribució de càrrega. No obstant, contràriament al client-servidor, les xarxes P2P requereixen solucions sofisticades a tots els nivells, que varien des de l'organització de la xarxa a la localització de recursos i la seua gestió. En aquesta tesi s'adreça un dels problemes clau de les xarxes P2P: ser capaç de realitzar eficientment cerques de recursos de naturalesa semàntica sota condicions realistes i dinàmiques. Existeixen nombroses solucions a aquest tema basades en la computació simple, evolutiva i també basades en l'estimèrgia (de l'anglès, stigmergy), però pocs esforços s'han realitzat per intentar resoldre l'ampli conjunt de reptes existent. En primer lloc, les xarxes P2P reals són rarament estàtiques: els nodes es connecten, desconnecten i canvien els seus continguts. A més a més, la incorporació trivial de cerques semàntiques als algorismes existents causa una disminució significant de l'eficiència de la cerca. En aquesta recerca s'ha construït una solució incremental, començant pel sistema clàssic de colònia de formigues (de l'anglés, Ant Colony System o ACS) dins de la metaheurística d'optimització de colònies de formigues (de l'anglès, Ant Colony Optimization o ACO). ACO és un entorn algorísmic utilitzat per cercar en grafs i que aborda el problema de forma satisfactòria, tot i que no proveeix d'una solució immediata a cap dels problemes anteriorment mencionats. Primer, s'ha proposat una variant eficient d'ACS en xarxes P2P estructurades (en forma d'hipercub) a través d'un algorisme de processament post-camí el qual s'ha anomenat en anglès Tabu Route Optimization (TRO). A continuació, s'ha procedit a resoldre el problema del dinamisme de les xarxes amb un enfocament de difusió d'informació compatible amb ACO. Com a conseqüència, s'ha intentat incorporar la component semàntica de les cerques. Aquest enfocament inicial al problema s'ha realitzat permetent a ACS diferenciar entre tipus de cerques amb la idea de ''feromona per concepte'', i s'ha anomenat a aquest producte Routing Concept ACS o RC-ACS. RC-ACS és una implementació multi-feromona robusta i estàtica d'ACS. No obstant, s'ha pogut concloure que l'enfocament de feromona per concepte ofereix només una escalabilitat limitada i no pot ser considerada una solució global. En aquest respecte s'ha realitzat progrés posteriorment introduint una nova idea a RC-ACS: la creació dinàmica de feromones, la qual reemplaça a l'assignació un a un de les mateixes. A l'algorisme resultant se l'ha anomenat en anglès Angry Ant Framework (AAF). En AAF es creen nous nivells de feromones a mesura que es necessiten durant la cerca, i no abans d'aquesta. El progrés final s'ha aconseguit quan s'ha permès a AAF, no sols crear nivells de feromones, sinó reassignar-los per optimitzar la utilització de feromones. L'algorisme resultant s'ha anomenat EntropicAAF i ha sigut avaluat com un dels algorismes per a cerques semàntiques P2P amb millors prestacions.<br>Krynicki, KK. (2016). Ant Colony Algorithms for the Resolution of Semantic Searches in P2P Networks [Tesis doctoral no publicada]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/61293<br>TESIS<br>Premiado

Hlavním cílem této práce je odhadnout výkonnost volně šířeni balík  Sémantický Vektory a třída MoreLikeThis z balíku Apache Lucene. Tato práce nabízí porovnání těchto dvou přístupů a zavádí metody, které mohou vést ke zlepšení kvality vyhledávání.

Thesis (M.S.)--University of Kentucky, 2003.<br>Title from document title page (viewed June 21, 2004). Document formatted into pages; contains viii, 55 p. : ill. Includes abstract and vita. Includes bibliographical references (p. 52-54).

We study the properties of a number of algorithms proposed to explore the computational space generated by a very simple and general idea: the notion of a mathematical definition and a number of suggested formal interpretations ofthis idea. Theories of partial inductive definitions (PID) constitute a class of logics based on the notion of an inductive definition. Formal systems based on this notion can be used to generalize Horn-logic and naturally allow and suggest extensions which differ in interesting ways from generalizations based on first order predicate calculus. E.g. the notion of completion generated by a calculus of PID and the resulting notion of negation is completely natural and does not require externally motivated procedures such as "negation as failure". For this reason, computational issues arising in these calculi deserve closer inspection. This work discuss a number of finitary theories of PID and analyzethe algorithmic and semantical issues that arise in each of them. There has been significant work on implementing logic programming languages in this setting and we briefly present the programming language and knowledge modelling tool GCLA II in which many of the computational prob-lems discussed arise naturally in practice.<br><p>Also published as SICS Dissertation no. SICS-D-19</p>

Peer-to-Peer(P2P) systems have emerged as a promising paradigm to structure large scale distributed systems. They provide a robust, scalable and decentralized way to share and publish data.The unstructured P2P systems have gained much popularity in recent years for their wide applicability and simplicity. However efficient resource discovery remains a fundamental challenge for unstructured P2P networks due to the lack of a network structure. To effectively harness the power of unstructured P2P systems, the challenges in distributed knowledge management and information search need to be overcome. Current attempts to solve the problems pertaining to knowledge management and search have focused on simple term based routing indices and keyword search queries. Many P2P resource discovery applications will require more complex query functionality, as users will publish semantically rich data and need efficiently content location algorithms that find target content at moderate cost. Therefore, effective knowledge and data management techniques and search tools for information retrieval are imperative and lasting. In my dissertation, I present a suite of protocols that assist in efficient content location and knowledge management in unstructured Peer-to-Peer overlays. The basis of these schemes is their ability to learn from past peer interactions and increasing their performance with time.My work aims to provide effective and bandwidth-efficient searching and data sharing in unstructured P2P environments. A suite of algorithms which provide peers in unstructured P2P overlays with the state necessary in order to efficiently locate, disseminate and replicate objects is presented. Also, Existing approaches to federated search are adapted and new methods are developed for semantic knowledge representation, resource selection, and knowledge evolution for efficient search in dynamic and distributed P2P network environments. Furthermore,autonomous and decentralized algorithms that reorganizes an unstructured network topology into a one with desired search-enhancing properties are proposed in a network evolution model to facilitate effective and efficient semantic search in dynamic environments.

<p>People use web search engines to fill a wide variety of navigational, informational and transactional needs. However, current major search engines on the web retrieve a large number of documents of which only a small fraction are relevant to the user query. The user then has to manually search for relevant documents by traversing a topic hierarchy, into which a collection is categorised. As more information becomes available, it becomes a time consuming task to search for required relevant information.</p> <p>This research develops an effective tool, the web document clustering (WDC) system, to cluster, and then rank, the output data obtained from queries submitted to a search engine, into three pre-defined fuzzy clusters. Namely closely related, related and not related. Documents in closely related and related documents are ranked based on their context.</p> <p>The WDC output has been compared against document clustering results from the Google, Vivisimo and Dogpile systems as these where considered the best at the fourth Search Engine Awards [24]. Test data was from standard document sets, such as the TREC-8 [118] data files and the Iris database [38], or 3 from test text retrieval tasks, "Latex", "Genetic Algorithms" and "Evolutionary Algorithms". Our proposed system had as good as, or better results, than that obtained by these other systems. We have shown that the proposed system can effectively and efficiently locate closely related, related and not related, documents among the retrieved document set for queries submitted to a search engine.</p> <p>We developed a methodology to supply the user with a list of keywords filtered from the initial search result set to further refine the search. Again we tested our clustering results against the Google, Vivisimo and Dogpile systems. In all cases we have found that our WDC performs as well as, or better than these systems.</p> <p>The contributions of this research are:</p> <ol> <li>A post-retrieval fuzzy document clustering algorithm that groups documents into closely related, related and not related clusters. This algorithm uses modified fuzzy c-means (FCM) algorithm to cluter documents into predefined intelligent fuzzy clusters and this approach has not been used before.</li> <li>The fuzzy WDC system satisfies the user's information need as far as possible by allowing the user to reformulate the initial query. The system prepares an initial word list by selecting a few characteristics terms of high frequency from the first twenty documents in the initial search engine output. The user is then able to use these terms to input a secondary query. The WDC system then creates a second word list, or the context of the user query (COQ), from the closely related documents to provide training data to refine the search. Documents containing words with high frequency from the training list, based on a pre-defined threshold value, are then presented to the user to refine the search by reformulating the query. In this way the context of the user query is built, enabling the user to learn from the keyword list. This approach is not available in current search engine technology.</li> <li>A number of modifications were made to the FCM algorithm to improve its performance in web document clustering. A factor sw_kq is introduced into the membership function as a measure of the amount of overlaping between the components of the feature vector and the cluster prototype. As the FCM algorithm is greatly affected by the values used to initialise the components of cluster prototypes a machine learning approach, using an Evolutionary Algorithm, was used to resolve the initialisation problem.</li> <li>Experimental results indicate that the WDC system outperformed Google, Dogpile and the Vivisimo search engines. The post-retrieval fuzzy web document clustering algorithm designed in this research improves the precision of web searches and it also contributes to the knowledge of document retrieval using fuzzy logic.</li> <li>A relational data model was used to automatically store data output from the search engine off-line. This takes the processing of data of the Internet off-line, saving resources and making better use of the local CPU.</li> <li>This algorithm uses Latent Semantic Indexing (LSI) to rank documents in the closely related and related clusters. Using LSI to rank document is wellknown, however, we are the first to apply it in the context of ranking closely related documents by using COQ to form the term x document matrix in LSI, to obtain better ranking results.</li> <li>Adjustments based on document size are proposed for dealing with problems associated with varying document size in the retrieved documents and the effect this has on cluster analysis.</li> </ol>

As we use the Web for social networking, shopping, and news, we leave a personal trail. These days, linger over a Web page selling lamps, and they will turn up at the advertising margins as you move around the Internet, reminding you, tempting you to make that purchase. Search engines such as Google can now look deep into the data on the Web to pull out instances of the words you are looking for. And there are pages that collect and assess information to give you a snapshot of changing political opinion. These are just basic examples of the growth of "Web intelligence", as increasingly sophisticated algorithms operate on the vast and growing amount of data on the Web, sifting, selecting, comparing, aggregating, correcting; following simple but powerful rules to decide what matters. While original optimism for Artificial Intelligence declined, this new kind of machine intelligence is emerging as the Web grows ever larger and more interconnected. Gautam Shroff takes us on a journey through the computer science of search, natural language, text mining, machine learning, swarm computing, and semantic reasoning, from Watson to self-driving cars. This machine intelligence may even mimic at a basic level what happens in the brain.

AbstractIn recent years the digital landscape has been rapidly evolving as the application of artificial intelligence (AI) becomes increasingly important in shaping search engine optimization (SEO) strategies and revolutionizing the way websites are optimized for search engines. This research aims to explore the influence of AI in the field of SEO through a literature review that is conducted using the PRISMA framework. The study delves into how AI capabilities such as generative AI and natural language processing (NLP) are leveraged to boost SEO. These techniques in turn allow search engines to provide more accurate, user-centric results, highlighting the importance of semantic search, where search engines understand the context and intent of a user’s search query, ensuring a more personalized and effective search experience. On the other hand, AI and its tools are used by digital marketers to implement SEO strategies such as automatic keyword research, content optimization, and backlink analysis. The automation offered by AI not only enhances efficiency but also heralds a new era of precision in SEO strategy. The application of AI in SEO paves the way for more targeted SEO campaigns that attract more organic visits to business websites. However, relying on AI in SEO also poses challenges and considerations. The evolving nature of AI algorithms requires constant adaptation by businesses and SEO professionals, while the black-box nature of these algorithms can lead to the opaque and unpredictable evolution of SEO results. Furthermore, the power of AI to shape online content and visibility raises questions about equality, control, and manipulation in the digital environment. The insights gained from this study could inform future developments in SEO strategies, ensuring a more robust, fair, and user-centric digital search landscape.

AbstractThe chapter provides a broad introduction to the foundations of health AI and ML systems and is organized as follows: (1) Theoretical properties and formal vs. heuristic systems: computability, incompleteness theorem, space and time complexity, exact vs. asymptotic complexity, complexity classes and how to establish complexity of problems even in the absence of known algorithms that solve them, problem complexity vs. algorithm and program complexity, and various other properties. Moreover, we discuss the practical implications of complexity for system tractability, the folly of expecting Moore’s Law and large-scale computing to solve intractable problems, and common techniques for creating tractable systems that operate in intractable problem spaces. We also discuss the distinction between heuristic and formal systems and show that they exist on a continuum rather than in separate spaces. (2) Foundations of AI including logics and logic based systems (rule based systems, semantic networks, planning systems search, NLP parsers), symbolic vs. non-symbolic AI, Reasoning with Uncertainty, Decision Making theory, Bayesian Networks, and AI/ML programming languages. (3) Foundations of Computational Learning Theory: ML as search, ML as geometrical construction and function optimization, role of inductive biases, PAC learning, VC dimension, Theory of Feature Selection, Theory of Causal Discovery. Optimal Bayes Classifier, No Free Lunch Theorems, Universal Function Approximation, generative vs. discriminative models; Bias-Variance Decomposition of error and essential concepts of mathematical statistics.

Syntax-guided synthesis is commonly used to generate programs encoding policies. In this approach, the set of programs, that can be written in a domain-specific language defines the search space, and an algorithm searches within this space for programs that encode strong policies. In this paper, we propose an alternative method for synthesizing programmatic policies, where we search within an approximation of the language's semantic space. We hypothesized that searching in semantic spaces is more sample-efficient compared to syntax-based spaces. Our rationale is that the search is more efficient if the algorithm evaluates different agent behaviors as it searches through the space, a feature often missing in syntax-based spaces. This is because small changes in the syntax of a program often do not result in different agent behaviors. We define semantic spaces by learning a library of programs that present different agent behaviors. Then, we approximate the semantic space by defining a neighborhood function for local search algorithms, where we replace parts of the current candidate program with programs from the library. We evaluated our hypothesis in a real-time strategy game called MicroRTS. Empirical results support our hypothesis that searching in semantic spaces can be more sample-efficient than searching in syntax-based spaces.

The problem of finding the most relevant documents as a result of an extended and refined query is considered. For this, a search model and a text preprocessing mechanism are proposed, as well as the joint use of a search engine and a neural network model built on the basis of an index using word2vec algorithms to generate an extended query with synonyms and refine search results based on a selection of similar documents in a digital semantic library. The paper investigates the construction of a vector representation of documents based on paragraphs in relation to the data array of the digital semantic library LibMeta. Each piece of text is labeled. Both the whole document and its separate parts can be marked. The problem of enriching user queries with synonyms was solved, then when building a search model together with word2vec algorithms, an approach of "indexing first, then training" was used to cover more information and give more accurate search results.

The problem of finding the most relevant documents as a result of an extended and refined query is considered. To solve it, a search model and a text preprocessing mechanism are proposed. It is proposed to use a search engine and a model based on an index using word2vec algorithms to generate an extended query with synonyms. To refine the search results, the idea of selecting similar documents in the digital semantic library is used. The paper investigates the construction of a vector representation of documents in relation to the data array of the digital semantic library LibMeta. Each piece of text is labeled. Both the whole document and its separate parts can be marked. Search through the library content, search for new terms and new semantic relationships between terms of the subject area becomes more meaningful and accurate. The task of enriching user queries with synonyms was solved. When building a search model in conjunction with word2vec algorithms, a "indexing first, then learning" approach is used, which allows obtaining more accurate search results. This work can be considered one of the first stages in the formation of a training data array for the subject area of problems of mathematical physics and the formation of a dictionary of synonyms for this subject area. The model was trained on the basis of the library's mathematical content. Examples of training, extended query and search quality assessment using training and synonyms are given.

This paper presents the final results of the research project that aimed to build a Semantic Search Engine that uses an Ontology and a model trained with Machine Learning to support the semantic search of research projects of the System of Research from the University of Nariño. For the construction of FENIX, as this Engine is called, it was used a methodology that includes the stages: appropriation of knowledge, installation and configuration of tools, libraries and technologies, collection, extraction and preparation of research projects, design and development of the Semantic Search Engine. The main results of the work were three: a) the complete construction of the Ontology with classes, object properties (predicates), data properties (attributes) and individuals (instances) in Protegé, SPARQL queries with Apache Jena Fuseki and the respective coding with Owlready2 using Jupyter Notebook with Python within the virtual environment of anaconda; b) the successful training of the model for which Machine Learning algorithms and specifically Natural Language Processing algorithms were used such as: SpaCy, NLTK, Word2vec and Doc2vec, this was also done in Jupyter Notebook with Python within the virtual environment of anaconda and with Elasticsearch; and c) the creation of FENIX managing and unifying the queries for the Ontology and for the Machine Learning model. The tests showed that FENIX was successful in all the searches that were carried out because its results were satisfactory.

Many algorithms for Knowledge-Based Question Answering (KBQA) depend on semantic parsing, which translates a question to its logical form. When only weak supervision is provided, it is usually necessary to search valid logical forms for model training. However, a complex question typically involves a huge search space, which creates two main problems: 1) the solutions limited by computation time and memory usually reduce the success rate of the search, and 2) spurious logical forms in the search results degrade the quality of training data. These two problems lead to a poorly-trained semantic parsing model. In this work, we propose an effective search method for weakly supervised KBQA based on operator prediction for questions. With search space constrained by predicted operators, sufficient search paths can be explored, more valid logical forms can be derived, and operators possibly causing spurious logical forms can be avoided. As a result, a larger proportion of questions in a weakly supervised training set are equipped with logical forms, and fewer spurious logical forms are generated. Such high-quality training data directly contributes to a better semantic parsing model. Experimental results on one of the largest KBQA datasets (i.e., CSQA) verify the effectiveness of our approach and deliver a new state-of-the-art performance.

The COVID-19 pandemic was a global health crisis that lasted until May 4, 2023, affecting millions of people and raising many questions about transmission, diagnosis, treatment, vaccine development, and viral pathogens. Unfortunately, misinformation created more socioeconomic damage than the disease itself. To address this problem, we have developed Cognitive Search, a user-friendly application service that uses the latest advances in Natural Language Processing (NLP) to retrieve information from CORD-19, a resource for scholarly articles on COVID-19 and related pathogens. This system uses a combination of Term-Frequency, Semantic Neural Research, and Hybrid Term-Neural algorithms to improve document retrieval performance. The Hybrid Term-Neural approach also considers temporal information in documents to provide more accurate search results. With an intuitive interface, this application can generate valuable insights to help combat outbreaks.