Search

Over the last decade, service selection and recommendation had been two strongly related service filtering steps. The ever changing services environment, users tastes, as well as the perception and popularity of available services, rise a question regarding the appropriate means to capture and analyze such changes over time. Most service recommendation solutions are static and do not offer a multi-relational modeling of user-service interactions over time. Time is a contextual dimension that has, recently, received a lot of attention, leading to a new class of recommender systems, called time-aware recommender systems. In this work, we propose a service recommendation method that takes advantage of temporal knowledge graphs. As a de facto standard to model multiple and complex interactions between heterogeneous entities, knowledge graphs will serve as a historical knowledge base for our TASR system. We, first, model the user-service interactions over time, by constructing a temporal service knowledge graph (TSKG) that will be later enriched through a completion step. Second, to explore the TSKG and extract top-rated services, we use Convolutional Neural Networks (CNN) to embed the TSKG into a low-dimensional vector space, facilitating then its mining.

The complex traffic and road environment pose considerable challenges to the accuracy, timeliness, and adaptive ability of connected and autonomous vehicles (CAVs) in making driving decisions. This paper uses vehicle collaboration and integrates the adaptive learning capabilities of machine learning and the interpretation capabilities of expert systems (ESs) in a unified architecture to form a hybrid autonomous driving guidance system, which not only solves the ``bottleneck'' of knowledge acquisition during the construction of expert systems but also solves the ``black box'' phenomenon of machine learning in the decision-making process. First, an autonomous driving strategy based on deep reinforcement learning (DRL) is proposed for CAVs to make decisions and extract corresponding rules. Next, we design an ES knowledge base expansion method including rule extraction, rule sharing, and rule test. Particularly, vehicular blockchain is adopted to ensure user privacy and data security during the rule-sharing process. Third, hybrid autonomous driving guidance combining ES and machine learning is proposed for CAVs to make accurate and efficient decisions in different driving environments. Once the strategy is well trained, it can effectively guide CAVs to cope with the complex traffic environment. Extensive simulations validate the performance of our proposal in terms of decision-making accuracy, effectiveness, and safety.

Fish classification (FC) is an expansively studied problem in the domains of image segmentation, pattern recognition, and information retrieval. It has been applied in a countless number of domains including target marketing. Meanwhile, governments are obliged to maintain the fish supply and balance between the ecosystem, commercial, agriculture field, marine scientists, and industrial arena of fish including the nutrition and canning factories. The various FC techniques performance is compared relying on the availability of preprocessing and feature extraction methods, the number of extracted features and classification accuracy, the number of fish families/species recognized. This survey also reviewed the use of Databases such as Fish4-Knowledge (F4K), knowledge database, and Global Information System (GIS) on Fishes and other FC databases. The study on preprocessing methods features extraction techniques and classifiers are gathered from recent works to enhance the understanding of the characteristics of preprocessing methods, features extraction techniques, and classifiers to guide future research directions and compensate for current research gaps.

The predictive vessel surveillance is one of the indispensable functional components in intelligent maritime traffic system. Vessel trajectory prediction serves as a prerequisite for collision detection and risk assessment. Perceiving the forthcoming traffic situation in advance helps decide the succeeding actions to mitigate the potential risk. The availability of maritime big data brings great potential to extract vessel movement patterns to support trajectory forecasting. In this paper, a novel vessel trajectory and navigating state prediction methodology is proposed based on AIS data, which synergizes properly designed learning, motion modelling and knowledge base assisted particle filtering processes. The primary contributions of this work also comprise several critical research findings to handle the key challenges in vessel trajectory and navigating state prediction problem, such as the adaptive training window determination for the learning process and effective knowledge storage and searching algorithm intended to reduce the query time of waterway pattern retrieval. The studies for these challenges are still missing in the reported literatures but they are essentially important for improving the prediction accuracy, efficiency and practicality. With the maritime traffic data collected for Singapore water, a thorough evaluation of the prediction performance has been conducted for different navigating scenarios. It is also observed that better prediction outperforms on account of allowing earlier alert in risk detection.

Traditional topic classification usually adopts the closed-world assumption that all the test topics have been seen in training. However, in open dynamic environments, the potential new topics may appear in testing due to the evolution of text data over time. Considering the uncertainty and multi-granularity of dynamic text data, such open topic classification needs to detect unseen topics by mining the boundary region continually, and incrementally update the previous models by knowledge accumulation. To address these challenge issues, this paper introduces a unified framework of three-way multi-granularity learning to open topic classification based on the fusion of three-way decision and granular computing. First, we propose the multilevel granular structure of tasks from the temporal-spatial multi-granularity perspective. Then, we construct an adaptive decision boundary and use the centroids and the corresponding radius to discover unknowns by the reject option. Subsequently, we further explore the unknown topics by three-way enhanced clustering and the uncertain instances will be re-investigated in the next stage. Besides, we design a built-in knowledge base represented as the centroid of each topic to store the topic knowledge. Finally, the experiments are conducted to compare the performances of proposed models and the efficiency of knowledge accumulation with classic models.

Higher education management problems in delivering 100% of graduates who can satisfy business demands. In industry it is often difficult for qualified graduates to identify the appropriate means to evaluate problem - solving abilities as well as shortcomings in the evaluation of problem solving skills. This is partially due to the lack of an adequate methodology. The purpose of this paper is to provide the appropriate CBR-KBS model for predicting and evaluating the characteristics of the student's dataset so as to comply with the parameters of selection required by the university industry. Machine learning algorithms have been used in these study areas under supervision, uncompleted and uncontrolled; K-Nearest neighbor, Naïve Bayes, Decision Tree, Neural Network, Logistic Regression and Vector Support Machines. The proposed model would allow university management to make easier, more professional, experienced and industry-specific plans for the manufacturing of graduates and graduates who passed the type I and II examinations held by the employment opportunities.

Automatic International Classification of Diseases (ICD) coding is an important task in the future of artificial intelligence healthcare. In recent years, a lot of traditional machine learning-based methods have been proposed, and they achieved good results on this task. However, these traditional machine learning-based methods for automatic ICD coding only focus on the semantic features of clinical notes and ignore the feature extraction of ICD titles that are the descriptions of ICD codes. In this paper, we propose a knowledge attention-based deep learning framework called KAICD for automatic ICD coding. KAICD makes full use of the clinic notes and the ICD titles. The semantic features of clinic notes are extracted by a multi-scale convolutional neural network. For ICD titles, we use attention-based Bidirectional Gated Recurrent Unit (Bi-GRU) to build a knowledge database, which can offer additional information. Depending on input clinic notes, we can use the attention mechanism to obtain different knowledge vectors from the knowledge database where some ICD titles are more relevant to the input clinic notes. Last, we concatenate the knowledge vectors and the semantic features of clinic notes, and use them for the final prediction. KAICD is tested on a public dataset Medical Information Mart for Intensive Care III (MIMIC III); it achieves micro-precision of 0.502, micro-recall of 0.428, and micro-f1 of 0.462, which outperforms other competing methods. Furthermore, the results of the ablation study show that the knowledge database of ICD titles learned by the attention-based Bi-GRU enhances the feature expression and improves the prediction performance.

Data driven method based on big data and deep model is an effective tool to solve the problem of online power system transient stability assessment. Nevertheless, when the system suffers from some large disturbances such as the changes of topology, there is no proper way for a model to supplement the original knowledge base but to retrain it from scratch. Meanwhile, with the continuous emergence of new scenarios, data management has become a new dilemma reflected in the increase of data accumulation and retraining time. To address these issues, a novel TSA framework with continual learning ability is proposed. First, the improved convolutional neural network-based orthogonal weight modification algorithm is selected as a transient stability predictor. Then, large disturbance is regarded as a new recognition task which can be continuously learned by the predictor. Through the proposed updating process, the original knowledge base can be updated and supplemented only by the data corresponding to new scenario. The test results on two power systems show that the proposed scheme can realize the tradeoff between the integrity of knowledge base and the update speed without losing too much accuracy in a limited capacity.

Active shooting (AS) violence poses a serious threat to public safety. Most importantly, this number is increasing at an alarming rate over the last few years. Unfortunately, the knowledge base on attack detection and civilian response strategy under an AS incident are very limited. Such a lack of understanding hinders the development of effective training modules, response, and mitigation strategies (for both civilians and law enforcement officers) for an event that is highly unpredictable and unfolds quickly (typically in less than 5 min). To fill this gap in the literature, this article proposes an innovative mathematical model which captures several important features (e.g., capacity of the facility and individual choices, heterogeneity of individual behavioral and choice sets, restriction on choice sets depending on the location of the shooter and facility orientation, and many others) which are essential for appropriately characterizing and analyzing the response strategy for civilians under an AS exposed environment. We demonstrate the applicability of the proposed model by implementing the effectiveness of the RUN.HIDE.FIGHT. program in an academic environment. Results indicate that cognitive delay, the initial distribution of the civilians, and building configurations (e.g., location and number of entrances/exits) play a significant role in the safety of civilians under an AS incident.

One of the algorithms stored in natural intelligence is the writing of Arabic numerals in Indonesian words. Algorithms in naturals intelligence are not easy to find. This problem gave us an idea to create artificial intelligence that tries to mimic natural intelligence algorithms. The proposed algorithm for building artificial intelligence is an R-Z rule-based system. This rule-based system contains a knowledge base of R-Z rules and a knowledge base of facts. In the knowledge base, the R-Z rule provides the R rule and the Z rule, while the facts knowledge base provides facts in the form of a definite standard number and an affix word. R-Z rule-based system for reasoning writing Arabic numerals in Indonesian words uses forward chaining. Artificial intelligence designs that mimic naturals intelligence in writing numbers in Indonesian words were made in C using Borland C++ 5.02 software. The experimental results show that by applying the R's rule of seven rules and Z's of twenty-five rules, the R-Z rule-based system can write Arabic numerals in Indonesian words from Arabic numerals "0" to Arabic numerals "9999999". For example, to write the Arabic number "10" in Indonesian words, the R-Z rule-based system starts with the R2 rule. Rule R2 takes action on Z3 to create new facts about Arabic numerals in the Indonesian word, namely "SEPULUH."

Cardiac arrhythmias are not life-threatening straight away but can cause serious heart-related complications if not medically handled appropriately. An electrocardiogram (ECG) captures the heart’s electric activity and has widespread usage due to its easy deployment and non-invasive aspect. Arrhythmia classification through manual analysis of electrocardiogram (ECG) is troublesome, tedious, and prone to human errors that can lead to serious repercussions. Hence, it’s a more effective alternative to deploy computational techniques to automatically perform the classification. Traditional techniques are data-driven and require an immense amount of data to train and then perform identification. This paper presents an ontology-driven knowledge model to automatically diagnose arrhythmias based on the patient’s sensor-based ECG data. The proposed approach models the arrhythmia domain knowledge and the conceptual relationships relevant to classification of heartbeat into corresponding cardiac arrhythmias and to facilitate decision making with respect to the patients. The newly developed arrhythmia ontology consists of three different modules, each semantically annotating a different aspect of the arrhythmia detection process. A SWRL (Semantic Web Rule Language) based ontology classifier performs classification of patient’s ECG data into corresponding cardiac arrhythmia types. The constructed knowledge base is ontologically aligned with some benchmarked top-level ontologies that promotes the semantic interoperability across multiple domains. The resultant ontological model is validated with a real-world ECG dataset and compared with the existing approaches showing higher precision rate and comparable performance. The developed model establishes a standardized ontology, that promotes the exchange and shareability of consensual domain knowledge about arrhythmic conditions, supports information retrieval and knowledge discovery.

Word embedding models have recently shown some capability to encode hierarchical information that exists in textual data. However, such models do not explicitly encode the hierarchical structure that exists among words. In this work, we propose a method to learn hierarchical word embeddings (HWEs) in a specific order to encode the hierarchical information of a knowledge base (KB) in a vector space. To learn the word embeddings, our proposed method considers not only the hypernym relations that exist between words in a KB but also contextual information in a text corpus. The experimental results on various applications, such as supervised and unsupervised hypernymy detection, graded lexical entailment prediction, hierarchical path prediction, and word reconstruction tasks, show the ability of the proposed method to encode the hierarchy. Moreover, the proposed method outperforms previously proposed methods for learning nonspecialised, hypernym-specific, and hierarchical word embeddings on multiple benchmarks.

This article presents the creation of an ontology describing Bulgarian folklore. The ontology has been developed based on the CCO (Cataloging Cultural Objects) standard. CCO leads us to provide data content standards. The set of ontologies is a knowledge base used by intelligent agents. One of the main problems with systems and developments describing different cultural and/or historical sites is the difficulty in sharing information between individual societies or systems. The use of standards greatly facilitates the dissemination of data. The CCO standard provides clear and precise definitions of the individual attributes that an object must possess. The ontologies developed by Protégé are designed in such a way as to meet the requirements of the standard. The created ontologies describe 7 folklore areas with specific ones being: music, dances and musical instruments. The work on creating this ontology directs us to create web applications of the developed ontologies.

Undoubtedly, the factor leading to the death of cattle is the spread of diseases that can worsenthe health of animals. Diagnosing animal diseases is difficult if a qualified veterinary professional is notavailable, which is usually due to the remoteness of livestock farms from scientific veterinary centres. Thus,the use of expert systems for solving veterinary medicine problems is proposed as a solution. An expertsystem for diagnosing animal diseases is required to simplify the work of veterinarians and subsequentlyreduce the incidence of animal diseases.The consequence is that the need to implement decision support toolsis undeniable, as veterinary medicine is a science that involves possible diagnostic errors, decision-makingcomplexity and an extensive knowledge base. As a result, the use of cognitive maps, the functionality ofwhich is described in this article, is proposed. The development of these decision support tools includesboth discrete analytical information and intuitive elements that optimise their impact on everyday veterinarypractice. Cognitive maps are active knowledge resources that use animal data to generate case-specificrecommendations to support diagnostic decision-making. The knowledge sources used and the process ofobtaining, analyzing and evaluating knowledge are described in detail. Productive rules are used as a toolto represent knowledge, which is also used to address the uncertainties that arise in the decision-makingprocess. Introducing one-dimensional tables when structuring a knowledge library is an appropriate solutionfor system design. Creating a knowledge base in the way described can solve the initial goals of developinga software product for veterinary applications.

Recently, pretrained language models, such as Bert and XLNet, have rapidly advanced the state of the art on many NLP tasks. They can model implicit semantic information between words in the text. However, it is solely at the token level without considering the background knowledge. Intuitively, background knowledge influences the efficacy of text understanding. Inspired by this, we focus on improving model pretraining by leveraging external knowledge. Different from recent research that optimizes pretraining models by knowledge masking strategies, we propose a simple but general method to transfer explicit knowledge with pretraining. To be specific, we first match knowledge facts from a knowledge base (KB) and then add a knowledge injunction layer to a transformer directly without changing its architecture. This study seeks to find the direct impact of explicit knowledge on model pretraining. We conduct experiments on 7 datasets using 5 knowledge bases in different downstream tasks. Our investigation reveals promising results in all the tasks. The experiment also verifies that domain-specific knowledge is superior to open-domain knowledge in domain-specific task, and different knowledge bases have different performances in different tasks.

International audience; Cooperative machine learning has many applications, such as data annotation, where an initial model trained with partially labeled data is used to predict labels for unseen data continuously. Predicted labels with a low confidence value are manually revised to allow the model to be retrained with the predicted and revised data. In this paper, we propose an alternative to this approach: an initial training process called Deep Unsupervised Active Learning. Using the proposed training scheme, a classification model can incrementally acquire new knowledge during the testing phase without manual guidance or correction of decision making. The training process consists of two stages: the first stage of supervised training using a classification model, and an unsupervised active learning stage during the test phase. The labels predicted during the test phase, with high confidence, are continuously used to extend the knowledge base of the model. To optimize the proposed method, the model must have a high initial recognition rate. To this end, we exploited the Visual Geometric Group (VGG16) pre-trained model applied to three datasets: Mathematical Image Analysis (AMI), University of Science and Technology Beijing (USTB2), and Annotated Web Ears (AWE). This approach achieved impressive performance that shows a significant improvement in the recognition rate of the USTB2 dataset by coloring its images using a Generative Adversarial Network (GAN). The obtained performances are interesting compared to the current methods: the recognition rates are 100.00%, 98.33%, and 51.25% for the USTB2, AMI, and AWE datasets, respectively.

Health is a blessing that is very important for humans, most of us, many of us don't really care about skin diseases caused by fungi such as tinea versicolor, water fleas, ringworm and scabies. In general, people are quite aware of how to deal with the symptoms of skin diseases suffered. But it would be better to include medical participation in detecting a disease symptom, because many disease symptoms are considered trivial by some people but can be fatal to human skin. So it is necessary to make an application based on medical knowledge to diagnose skin diseases in humans which is used as a tool in obtaining information about skin diseases in humans and provide recommendations as the first action that must be taken to repeat skin diseases in humans. The knowledge base is structured in such a way into a database with several tables. Drawing conclusions in this expert system using the forward chaining inference method. The expert system will provide questions to the user in the form of symptoms of several diseases and the user will answer these questions. Until the user will get a solution from the results of the question earlier.

Machine Reading Comprehension (MRC) is a challenging but meaningful task in natural language processing (NLP) that requires us to teach a machine to read and understand a given passage and answer questions related to that passage. In this paper, we present a rich knowledge-enhanced reader (RKE-Reader), a hierarchical MRC model that employs double knowledge bases with an NER system as its knowledge enhancement unit. Besides, we are the first to propose an offline answer-imporving method to help model to determine the uncertain answer without extra online training process. Our experimental results indicate that on most datasets, the RKE-Reader significantly outperforms most of the published models that do not have knowledge base, especially on datasets that need commonsense reasoning. And the ablation study also reflects that external knowledge bases and answer-selecting unit do make a positive contribution in the entire model.

With the increasing usage of information technology on the criminal side, the digital forensic analysis, especially multimedia forensics, becomes an emerging technique for cybercrime investigators to improve examination efficiency. The study focuses on the digital triage problem for evidence location during the automatic forensic process. After defining the multi-scale knowledge base for storing digital forensic investigators’ prior knowledge, a variable scale case-based reasoning method (VSCBR) is proposed to support investigators predicting evidential areas. The variable-scale clustering algorithm based on the scale transformation strategy (VSC-STS) is also put forward, which could identify highly similar past cases containing candidate evidence in the case reuse and revise phase. A case study is established using a real 15.9 GB bidding case dataset, which contains both text bidding documents and image technical drawings. Numerical experimental results show that the validation of the proposed VSC-STS is significantly improved compared with the traditional single-scale clustering algorithm, and it is insensitive to the initial parameter threshold. Moreover, the proposed method VSCBR is able to help investigators locate suspicious rule-violating evidences in practice.

Interdisciplinary approaches to creating high-tech computer systems for the diagnosis of melanoma using artificial intelligence are presented. A model is proposed for the architecture of an interactive expert system. This includes a set of features for a contemporary medical algorithm (the Kittler algorithm) along with a knowledge base and a diagnosis evaluation score for the case under study.

Candidate ranking is the process of selecting the candidate with the best matching probability to the question after generating candidates in the knowledge base question answering (KBQA) task. It is a representative problem in mining matching relationships between candidates and questions. Previous research works always model questions and candidate representations separately, ignoring their impact on each other. The text information is too short to capture rich features in the KBQA task. Therefore, our work presents an attention neural network (ANN) fused with structural features (SF-ANN) to rank candidate facts jointly. First, two types of attention mechanisms are used to capture the correlation between the question and the candidate fact: a mutual-attention mechanism that captures the correspondence between the sentence components of a question and each part of a candidate and an intra-attention mechanism that captures the self-dependency of the concatenation between a question and a candidate fact. Second, an ANN is designed for fusing these two types of attention mechanisms to deeply couple interactive information of the input. Finally, knowledge base structural features are introduced to supplement the semantic information to increase the richness of the information. Three mutual attention mechanisms are applied for fusing them into the ANN, resulting in higher information gain. The experimental results on the SimpleQuestions (SimpleQ) benchmark demonstrate that the proposed model achieves a higher ranking accuracy (82.9%) than the state-of-the-art models. Moreover, the ablation study on SimpleQ and WebQuestionsSP (WebQSP) shows that leveraged features and the propoesd ANN both contribute to performance improvement.

The word similarity task is used to calculate the similarity of any pair of words, and is a basic technology of natural language processing (NLP). The existing method is based on word embedding, which fails to capture polysemy and is greatly influenced by the quality of the corpus. In this paper, we propose a multi-prototype Chinese word representation model (MP-CWR) for word similarity based on synonym knowledge base, including knowledge representation module and word similarity module. For the first module, we propose a dual attention to combine semantic information for jointly learning word knowledge representation. The MP-CWR model utilizes the synonyms as prior knowledge to supplement the relationship between words, which is helpful to solve the challenge of semantic expression due to insufficient data. As for the word similarity module, we propose a multi-prototype representation for each word. Then we calculate and fuse the conceptual similarity of two words to obtain the final result. Finally, we verify the effectiveness of our model on three public data sets with other baseline models. In addition, the experiments also prove the stability and scalability of our MP-CWR model under different corpora.

The upstream oil and gas industry has been struggling with massive data files for decades. These monolithic files are supposed to be a valuable asset for companies, but in reality they are generally underutilized or even discarded. In this paper, we present the application of big data technique in the oil fields of western South China sea to predict the reservoir parameters and evaluate well productivity. The data files are firstly classified and stored by clustering analysis method on the basis of different labels. Combining different analysis tools, the variation law and correlations between different parameters are studied, and the knowledge base is thus established. We analyze the distribution laws of formation and fluid parameters, calculate the relationship between different parameters, and adopt the theoretical/empirical formula methods, to minimize the uncertainties of prediction results. After that, the oil productivities are comprehensively evaluated by clustering analysis and grey correlation methods. The productivity charts are built for all datasets and the classification rule according to sedimentary facies, reservoir member and formation oil state. Based on the clustering results, we analyze the key factors that impact oil production rate, and obtain the multi-parameter empirical productivity equations to evaluate well productivity. Finally, we establish the real-time-update software platform that integrates knowledge base, job management, parameter prediction, test design, interpretation, etc. Compared with the current methods of parameters’ prediction and productivity evaluation which need extensive experiment work or plenty of known input parameters, we integrate all available data sets and different approaches by big data technique to address the key uncertainties rather analyze each data source individually.

Automated text information extraction from cancer pathology reports is an active area of research to support national cancer surveillance. A well-known challenge is how to develop information extraction tools with robust performance across cancer registries. In this study we investigated whether transfer learning (TL) with a convolutional neural network (CNN) can facilitate cross-registry knowledge sharing. Specifically, we performed a series of experiments to determine whether a CNN trained with single-registry data is capable of transferring knowledge to another registry or whether developing a cross-registry knowledge database produces a more effective and generalizable model. Using data from two cancer registries and primary tumor site and topography as the information extraction task of interest, our study showed that TL results in 6.90% and 17.22% improvement of classification macro F-score over the baseline single-registry models. Detailed analysis illustrated that the observed improvement is evident in the low prevalence classes.

Knowledge bases allow data organization and exploration, making easier the data semantic understanding and its use by machines. Traditional strategies for knowledge base construction and augmentation have mostly relied on manual effort or automatic extraction of content from structured and semi-structured sources. In this work, we present DeepEx, a system that autonomously extracts missing attributes of entities in knowledge bases from unstructured text. We use Wikipedia as data source. Given entities on Wikipedia represented by their articles (text and infobox), DeepEx uses a classifier to detect sentences in the articles mentioning the possible missing attributes of the entities and then employs a deep-learning extraction model on those sentences to identify the attributes. The sentence classifier and attribute extractor are built with labels automatically produced by a weak supervision approach using infobox structured information as supervision source. We have compared our strategy with previous approaches to this problem on 29 different attributes from 4 domains. The results showed that our extraction pipeline achieved statistically superior performance in comparison with some baselines and variations of our approach.

This paper describes EPROP, a novel technique requiring little prior knowledge for word sense disambiguation of semantic relations between pairs of ambiguous concepts in knowledge bases. Our method makes inferences by aggregating evidences from ambiguous word interpretations and propagating the acquired knowledge over a taxonomy to generalize or specialize this knowledge. This propagation process allows the estimation of the degree of belief for each possible word sense assignment given the available evidence. EPROP only requires a sense inventory structured as a taxonomy to disambiguate a knowledge base by combining evidence from the ambiguous facts stored in the knowledge base. We have performed different experiments that show that our method achieves good results on the disambiguation of the semantic relations included in WordNet and ConceptNet. We also show how our method can be used to improve the performance of state-of-the-art word sense disambiguation methods.

Knowledge graphs (KGs) are one of the most common frameworks for knowledge representation. However, they suffer from a severe scalability problem that hinders their usage. KG embedding aims to provide a solution to this issue. Nonetheless, general approaches are incapable of representing and reasoning about information not previously contained in the graph. This paper proposes to leverage semantic and ontological information for a significant benefit of knowledge graph completion, focusing on triple classification. The goal of this task is to determine whether a given fact holds. Furthermore, this paper also considers the classification of facts that include entities that have not been seen during training, denoted out-of-knowledge-base or OOKB entities. An incremental method is presented, composed of six stages. Although the proposal can be applied to any KG embedding model, this work focuses on its application for semantic matching models, such as ComplEx and DistMult. Compared to other approaches, our proposal is model-agnostic, computationally inexpensive, and does not require retraining. The results show that triple classification accuracy scales up to 15% with the proposed approach, as well as accelerating the convergence of the model to its optimal solution. Furthermore, facts containing OOKB entities can be classified with a reasonable accuracy.

The distributed computing is huge gatherings of remote servers are organized to permit the incorporated information stockpiling. It has the access of computer services, resources and can be classified as public, private and reserved. In this study, we explored various types of Data bases used in cloud computing with respect to the category of Knowledge database, XML database, Online databases and Real-Time databases to improve the storage and data efficacy.

This article considers the problem of finding text documents similar in meaning in the corpus. We investigate a problem arising when developing applied intelligent information systems that is non-detection of a part of solutions by the TF-IDF algorithm: one can lose some document pairs that are similar according to human assessment, but receive a low similarity assessment from the program. A modification of the algorithm, with the replacement of the complete vocabulary with a vocabulary of specific terms is proposed. The addition of thesauri when building a corpus vector model based on a ranking function has not been previously investigated; the use of thesauri has so far been studied only to improve topic models. The purpose of this work is to improve the quality of the solution by minimizing the loss of its significant part and not adding “false similar” pairs of documents. The improvement is provided by the use of a vocabulary of specific terms extracted from the text of the analyzed documents when calculating the TF-IDF values for corpus vector representation. The experiment was carried out on two corpora of structured normative and technical documents united by a subject: state standards related to information technology and to the field of railways. The glossary of specific terms was compiled by automatic analysis of the text of the documents under consideration, and rule-based NER methods were used. It was demonstrated that the calculation of TF-IDF based on the terminology vocabulary gives more relevant results for the problem under study, which confirmed the hypothesis put forward. The proposed method is less dependent on the shortcomings of the text layer (such as recognition errors) than the calculation of the documents’ proximity using the complete vocabulary of the corpus. We determined the factors that can affect the quality of the decision: the way of compiling a terminology vocabulary, the choice of the range of n-grams for the vocabulary, the correctness of the wording of specific terms and the validity of their inclusion in the glossary of the document. The findings can be used to solve applied problems related to the search for documents that are close in meaning, such as semantic search, taking into account the subject area, corporate search in multi-user mode, detection of hidden plagiarism, identification of contradictions in a collection of documents, determination of novelty in documents when building a knowledge base.

The main aim of this paper is to design and test new software application to support experts involved in decision-making in the field of energy security. The prototype named ESecFuzzy is designed for this purpose. It implements various fuzzy concepts (fuzzy rules, fuzzified input data, fuzzy variables utilized by fuzzy terms and related fuzzy sets) chosen to obtain reasoning in case of unclear, uncertain or even incomplete input data for reaching a conclusion about them. The research was conducted on a sample of 28 European Union countries over a ten-year period (2008–2017). The research outcomes are as follows: (a) ESecFuzzy application for energy security measurement was developed; (b) application testing; (c) measuring energy security on a given sample, and (d) evaluation of the accuracy and acceptability of fuzzy logic as an approach for energy security quantification. ESecFuzzy provides the experimenting with different input data (time series intersected with geospatial data) as well as fine-tuning of fuzzy rules in knowledge base. Through the evaluation and experimenting process, ESecFuzzy application demonstrated the advantage of fuzzy logic by obtaining the derived information based on mass of heterogeneous data organized in time series that indicate general trends in the target application domain.

The paper is devoted to the development of instrumental and methodological support for the assessment of students' knowledge in mathematics in the framework of a hybrid intellectual learning environment. According to the structural scheme of the model of intellectual assessment of knowledge and creative potential of students in mathematics, the structure of the corresponding knowledge base is formalized. Knowledge base within the framework of data mining methods to create a module for monitoring and evaluating knowledge in mathematics has been developed. To model the educational process the authors propose methods for solving clustering, dimensionality reduction, and classification problems based on machine learning and artificial neural networks. The basics for the development of integrated knowledge base complexes and artificial intelligence tools for the educational process are presented. The obtained results are aimed at creating methods which ensure the processes of operational training, control and evaluation of knowledge.

In nowadays manufacturing, each technical assistance operation is digitally tracked. This results in a huge amount of textual data that can be exploited as a knowledge base to improve these operations. For instance, an ongoing problem can be addressed by retrieving potential solutions among the ones used to cope with similar problems during past operations. To be effective, most of the approaches for semantic textual similarity need to be supported by a structured semantic context (e.g. industry-specific ontology), resulting in high development and management costs. We overcome this limitation with a textual similarity approach featuring three functional modules. The data preparation module provides punctuation and stop-words removal, and word lemmatization. The pre-processed sentences undergo the sentence embedding module, based on Sentence-BERT (Bidirectional Encoder Representations from Transformers) and aimed at transforming the sentences into fixed-length vectors. Their cosine similarity is processed by the scoring module to match the expected similarity between the two original sentences. Finally, this similarity measure is employed to retrieve the most suitable recorded solutions for the ongoing problem. The effectiveness of the proposed approach is tested (i) against a state-of-the-art competitor and two well-known textual similarity approaches, and (ii) with two case studies, i.e. private company technical assistance reports and a benchmark dataset for semantic textual similarity. With respect to the state-of-the-art, the proposed approach results in comparable retrieval performance and significantly lower management cost: 30-min questionnaires are sufficient to obtain the semantic context knowledge to be injected into our textual search engine.

With the rapid development of information technology, English education has attracted the interest of scholars for its ability to rely on computers to analyze and understand human language. Taking machine learning (ML) as the core, this paper tries to develop an evaluation system for ML-based English education. Specifically, a basic ML model was established with four stages: environment, knowledge base, learning process, and implementation process. The entire ML system was divided from top to bottom into a user layer, a business layer, and a data layer. Application results show that, during the ML, even users with similar personal data and the same goal have vastly different suitable learning materials, due to their gap in personal capabilities. The research provides an effective way to evaluate English education in the context of computer science and artificial intelligence.

Motivation Medical terminology normalization aims to map the clinical mention to terminologies coming from a knowledge base, which plays an important role in analyzing electronic health record and many downstream tasks. In this article, we focus on Chinese procedure terminology normalization. The expressions of terminology are various and one medical mention may be linked to multiple terminologies. Existing studies based on learning to rank does not fully consider the quality of negative samples during model training and the importance of keywords in this domain-specific task. Results We propose a combined recall and rank framework to solve these problems. A pair-wise Bert model with deep metric learning is used to recall candidates. Previous methods either train Bert in a point-wise way or based on a multi-class classification problem, which may lead serious efficiency problems or not be effective enough. During model training, we design a novel online negative sampling algorithm to activate the pair-wise method. To deal with multi-implication scenarios, we train the task of implication number prediction together with the recall task in a multi-task learning setting, since these two tasks are highly complementary. In rank step, we propose a keywords attentive mechanism to focus on domain-specific information such as procedure sites and procedure types. Finally, a fusion block merges the results of the recall and the rank model. Detailed experimental analysis shows our proposed framework has a remarkable improvement on both performance and efficiency. Availability and implementation The source code will be available at https://github.com/sxthunder/CMTN upon publication.

Based on the dual-inheritance framework of cultural evolution, an improved multiobjective cultural algorithm (IMOCA) with a multistrategy knowledge base is presented in this paper. Inspired by the original versions of the cultural algorithm (CA), four basic types of knowledge sources, i.e., normative, situational, topographical and historical knowledge, are effectively utilized in the proposed IMOCA. Several modifications with the knowledge base of the IMOCA are made to tackle the characteristics of the multiobjective problem. Situational knowledge is used as an external repository for storing elite individuals, and the redesigned topographical knowledge functions as a search engine to broaden the expansion of the obtained solution set. The historical knowledge used in the IMOCA aims to select a productive knowledge source to generate new individuals. Furthermore, a simple mutation scheme is introduced into the knowledge base as an influence function for the purpose of fine tuning in the late stage of search. After configuring the parameters used in IMOCA, two classic benchmark suites, i.e., WFG and MaF, are used to assess the performance of the IMOCA in approaching the Pareto fronts (PFs) with accuracy and diversity. Nondominated solution sets obtained by the IMOCA are compared with 8 state-of-the-art multiobjective algorithms available in the literature. A statistical analysis is conducted, which reveals that, by modifying the basic knowledge structure of the CA, the proposed multiobjective cultural algorithm is competent enough to handle multiobjective problems with competitive performance.

Modern biomedical technologies development affords to provide the doctor with colossal amount of information about patients organism condition. However, the opportunity of using this data for medical diagnosis fully now is a distantive perspective only. The reason is a humans limited ability in assessment and interpretation this data arrays. The solution seems in artificial intelligence and expert systems wide introduction to medicine. Currently, almost all authors consider various options for constructing artificial neural networks as a way to implement artificial intelligence. This approach, which goes back to the fundamental theorem of A.N. Kolmogorov, the works of V.I. Arnold and Hecht-Nielsen [3], demonstrates excellent capabilities in a number of pattern recognition problems, which are reduced to revealing hidden details against the background of input noises. Much less often is mentioned such a method of modeling formal thinking as expert systems, which arose in the 1960s and then went into the shadows. Since the inception of cybernetics, computer programmers have tried to reproduce the mechanism of human thinking, that is, the task was to teach the computer to "think". The first known results in the field of creating and using intelligent systems were laid by the work of Norbert Wiener and G.S. Altshuller. At the same time, the creation of intelligent systems was reduced to the development of programs that solve problems using a variety of heuristic methods based on the property of human thinking to generalize.

There exist numerous resources online to gain the desired level of knowledge on any topic. However, this complicates the process of selecting the most appropriate resources. Every learner differs in terms of their learning speed, proficiency, and preferred mode of learning. This paper develops an adaptive learning management system to tackle this challenge. It creates a customized course for every student based on their level of knowledge, preferred mode of learning and continuously updates the course based on their learning speed. The material is filtered from a knowledge base that is dynamically updated using web scraping and ranked using feedback from students on the relevance and quality of each material. The model is tested in two phases: the content generation algorithm and the learnability of the system as a whole. The evaluation is done both quantitatively and qualitatively and validated with statistical analysis. Real-time testing of the system shows state-of-the-art performance.

PurposePrevious knowledge base question answering (KBQA) models only consider the monolingual scenario and cannot be directly extended to the cross-lingual scenario, in which the language of questions and that of knowledge base (KB) are different. Although a machine translation (MT) model can bridge the gap through translating questions to the language of KB, the noises of translated questions could accumulate and further sharply impair the final performance. Therefore, the authors propose a method to improve the robustness of KBQA models in the cross-lingual scenario.Design/methodology/approachThe authors propose a knowledge distillation-based robustness enhancement (KDRE) method. Specifically, first a monolingual model (teacher) is trained by ground truth (GT) data. Then to imitate the practical noises, a noise-generating model is designed to inject two types of noise into questions: general noise and translation-aware noise. Finally, the noisy questions are input into the student model. Meanwhile, the student model is jointly trained by GT data and distilled data, which are derived from the teacher when feeding GT questions.FindingsThe experimental results demonstrate that KDRE can improve the performance of models in the cross-lingual scenario. The performance of each module in KBQA model is improved by KDRE. The knowledge distillation (KD) and noise-generating model in the method can complementarily boost the robustness of models.Originality/valueThe authors first extend KBQA models from monolingual to cross-lingual scenario. Also, the authors first implement KD for KBQA to develop robust cross-lingual models.

Mining of regular trends in group action databases, time series databases, and lots of different database types was popularly studied in data processing research. Most previous studies follow the generation-and-test method of associate degree Apriori-like candidate collection. In this study, we seem to propose a particular frequency tree like structure, which is associated degree of prefix-tree like structure that is extended to be used for compressed storage, crucial knowledge of the frequency pattern, associated degrees create an economic FP-tree mining methodology, FP growth, by the growth of pattern fragments for the mining of the entire set of frequent patterns. Three different mining techniques are used to outsize the information which is compressed into small structures such as FP-tree that avoids repetitive information scans, cost. The proposed FP-tree-based mining receives an example philosophy of section creation to stay away from the exorbitant age of several competitor sets, and an apportioning-based, separating and-overcoming technique is used to divide the mining task into a contingent knowledge base for restricted mining designs that effectively reduces the investigation field.

With technical and industrial advances which have become a significant part of today's economic growth, remarkable advances have been made in the development and use of technical capabilities that enhance the endeavor of the medical profession. The developed countries have adopted the essential ''corner stone'' of database and registries development for monitoring birth defects. The use of National and Regional Registries has become ''a must'' to have a database that is necessary for rehabilitation, control and prevention programs. In Saudi Arabia, towards this objective, a study to establish a database and national registry to monitor birth defects and determine frequency, type, distribution and magnitude of the problem has been initiated in coordination with the Department of Human Genetics and the Computer department at the National Research Center aiming at developing an Arab Registry for Congenital and Genetic Malformation. The Prince Center for Disability Research (PSCDR)as well as the outpatient clinic for the human genetics department have developed a comprehensive research plan to study various aspects of handicaps and to assess the present care delivery status and consequently devise the most appropriate and applicable plan of care. Specific forms have been designed or collection of relevant information. The completeness and accuracy of reporting is ascertained by close supervision and adequate training of the team (composed of doctors, nurses and social workers) and a system by which uniformity of diagnosis can be curtained. It is expected that the data derived from this study will establish a regional registries of the handicapped, incorporating congenital and genetic anomalies in Arabs and will provide the basis for further research in prevention and care programs combating the handicap problem. Towards this goal, this paper uses a pediatric knowledge base acquired from medical geneticists and proposes a system containing mainly three modules a knowledge base module, a knowledge processing module, and an interface module. This endeavor concentrates mainly on the genetic symptoms knowledge base representation that allows the knowledge processing module to perform diagnosis advise according to the abnormal features and assist in detecting new syndromes. The system can be extended and gene or the medical problem domains.

Web service composition is a major way of constructing SOA-based applications. However, as uses’ requirements change, web services have to be recomposed correspondingly once again from the scratch. It will be rather time-consuming, error-prone and mostly fussy. To tackle the widespread requirements changes, we propose a novel approach that can make an existing composed service automatically grade to reach another new composed service in an evolutionary manner according to user’s requirements. An evolution model called control structure net is built to formally represent composition structure of a certain composed service based on interface dependence. Furthermore, a global dependence net, which provides an evolution knowledge base, is constructed by modeling all available web services. Evolution process is presented in detail and evolution reasoning algorithms are given to automatically remove invalid paths and make up necessary paths. Experimental results show that our proposed approach can correctly evolve to target composed service, and its performance also greatly surpasses that of classic service composition approach.

Tolerances and the buildup of tolerances affect the buildability, cost and quality of a product. The effect of the tolerances is so important that there are three major groups in manufacturing organization that are concerned with the tolerances which are design. process engineering and quality control. This work is a step in the field of computer-aided tolerancing (EXCAT). It consists of a helpful program which has a great important to the production and the design engineers. The program is written with an artificial intelligence programming language which is namely TURBO PROLOG, so the program is structured as mini expert system. For this reason the system contains a knowledge base for fit selections according to the function of the assembly, and limit gauges design. One of the benefits of using of the software is saving in the design and production time and so saving in their costs for the long run. This system is suitable for using in small workshop and also may belinked with CAD/CAM system.

One of the current trends in the development of the mobile application market is the development and adaptation of mobile applications that perform the functions of expert systems. A mobile app is software designed to work on tablets, smartphones, and other mobile devices. An expert system is a set of components that allow the system to make decisions at the level of a competent specialist (expert). The paper considers the concept of building knowledge base models based on a hierarchical functional system and its use in online consultation for mobile expert systems. The article analyzes the structure of knowledge base models in information and communication technologies for choosing: a mobile operator package, a provider, a hoster, and an online store. The model of the domain knowledge base in the “KARKAS” system consists of a hierarchy of domain classes, relations between them (rules) that operate within this model. The “KARKAS” system is a tool for developing knowledge base models for expert systems. To represent knowledge, a hierarchical functional system is developed, which is generated by the system based on the rules of products and frames. The output machine uses a hierarchical functional system during user consultation. The following modes of operation of the output machine are possible: direct output, reverse output, indirect output, Bayes formulas, criteria table, when the product sequence is a list of parameters. The system is implemented using the Embarcadero Delphi 10.4 platform. With the advent of the smartphone, the concept of an intelligent interface for messaging with a mobile application has increased in popularity. The TELEGRAM messenger chatbots are integrated into the “KARKAS” system: @Ribs_karkas_bot, @es_test_karkas_bot, @es_economy_karkas_bot, @es_info_tech_karkas_bot, which allow you to conduct online consultation, exchange messages between knowledge base models and chatbots. Embarcadero's cross-platform FireMonkey (FMX) framework allows you to quickly create application prototypes for running on different platforms: Windows, Android, and iOS. The purpose of the research is to develop and justify a fully functional prototype of the “KARKAS” tool system with a user interface in the form of a mobile application for the Android and iOS platforms. The problem of identifying temporal knowledge is urgent in solving many problems in the field of artificial intelligence. There are several ways to solve it, for example, the traditional direction is the explicit use of time in temporal knowledge models. Another approach is the use of time implicit in the idea of separation of the knowledge base. Functional system ‒ this is a system formed to achieve a given useful result (objective function) in the course of its functioning. Its system-forming factor is a specific result. In other words, the goal is considered as a given result, and the constraints are considered as the degree of freedom necessary to achieve the result. For example, a functional system can be considered as a set of functions with a certain set of operations applied to these functions. The role of functions is played by the rules of the knowledge base, and the main operations are the mapping of an attribute to a sample and determining the conditions for the applicability of the rules. The functional system is characterized by the following properties: connectivity-a chain of bundles of the knowledge base; complexity-hierarchy of levels of local knowledge bases; stability (adaptive behavior of the system) – the structure of the digraph of the functional system does not change with vertical perturbations of the rules. In other words, only the values of the antecedents of the rules for the local knowledge bases of the bundle chain are changed, and the bundle base, which is interpreted as the external environment, remains unchanged. In this paper, the knowledge base model is considered as a hierarchical functional system in which the result has an organizing influence on all stages of ontology formation. Classes and relationships between them can be considered as a logical construction of a functional system.

Internet of Things (IoT) brought great convenience to people’s daily lives. Meanwhile, the IoT devices are facing severe attacks from hackers and malicious attackers. Hackers and malicious attackers use various methods to invade the Internet of Things system, causing the Internet of Things to face a large number of targeted, concealed, and penetrating potential threats, which makes the privacy problem of the Internet of Things suffers serious challenges. But the existing methods and technologies cannot fully identify the attacker’s attack process and protect the privacy of the Internet of Things. Alarm correlation method can construct a complete attack scenario and identify the attacker’s intention by alarming the alarm data which provides an effective protection for user privacy. However, the existing alarm correlation methods still have the disadvantages of low correlation accuracy, poor correlation efficiency, and strong dependence on the knowledge base. To address these issues, we propose an alarm correlation method based on Affinity Propagation (AP) clustering algorithm and causal relationship. Our method considers that the alarm data triggered by the same attack process has high similarity characteristics, adopts the AP algorithm to improve the correlation efficiency, and at the same time constructs a complete attack process based on the causal correlation idea. The new alarm correlation method has a high correlation effect and builds a complete attack process to help managers identify attack intentions and prevent attacks.

Implementing third-party services to develop a chatbot is not cost-effective for many small and medium enterprises especially considering that many services only supports one language at a time. To address this gap, this paper proposes a multilingual chatbot system that will allow companies and organizations to customize and deploy their own multilingual chatbot service with two extended features, which are live chat and ticketing system. The chatbot will also be able to understand and reply in English, Malay, and Chinese as well as customizable given the dialogue shell and knowledge base. To achieve this, the development uses TypeScript for frontend web application while Go as the backend development. The development language for mobile application is Dart and the User Interface (UI) library is React. Finally, the database management system used is MongoDB. The developed prototype is then evaluated via a survey questionnaire and the findings suggested that the proposed system would be able to assist small and medium-sized business and organizations to deploy their own chatbot system as an alternative to existing customer services.

Diagnosa dini penyakit kulit yang disebabkan oleh infeksi jamur diperlukan untuk mengurangi resiko penyakit. Pada artikel ini, diusulkan sistem pakar dengan mengimplementasi teorema Bayes sebagai solusi diagnosa dini penyakit kulit, khususnya yang disebabkan oleh infeksi jamur. Akuisisi pengetahuan dari pakar, dalam hal ini dokter spesialis penyakit kulit dan kelamin, dilakukan untuk memperoleh basis pengetahuan. Sebanyak 31 data yang berasal dari rekam medis digunakan dalam penelitian ini. Jenis penyakit kulit yang dapat dideteksi yaitu tinea korporis, tinea unguium, pitiriasis versikolor, dan kandidosis mukokutan ringan. Teorema Bayes diimplementasikan pada mesin inferensi, sehingga menghasilkan kemungkinan jenis penyakit kulit yang diderita berdasarkan masukan gejala dari pengguna. Hasil dari sistem kemudian divalidasi oleh pakar. Berdasarkan hasil validasi tersebut, maka diperoleh 27 hasil yang sesuai dan 4 hasil yang tidak sesuai, sehingga akurasi kesesuaian sistem berdasarkan hasil tersebut adalah 87,1%.

Rule-based production expert systems have been used in medicine for a long time. The article shows how a mivar expert system can be used for the diagnosis of diabetes mellitus. Development of such a system is described. It involves creating a knowledge base in the form of a mivar net of parameters and rules, which is then entered in Wi!Mi to obtain the parameters and carry out the diagnosis. A prototype of the mivar knowledge base was created in the form of tables containing 62 parameters and 27 rules. In the future, evolution of the mivar knowledge base should lead to its growth.

Machine Learning has made impressive advances in many applications akin to human cognition for discernment. However, success has been limited in the areas of relational datasets, particularly for data with low volume, imbalanced groups, and mislabeled cases, with outputs that typically lack transparency and interpretability. The difficulties arise from the subtle overlapping and entanglement of functional and statistical relations at the source level. Hence, we have developed Pattern Discovery and Disentanglement System (PDD), which is able to discover explicit patterns from the data with various sizes, imbalanced groups, and screen out anomalies. We present herein four case studies on biomedical datasets to substantiate the efficacy of PDD. It improves prediction accuracy and facilitates transparent interpretation of discovered knowledge in an explicit representation framework PDD Knowledge Base that links the sources, the patterns, and individual patients. Hence, PDD promises broad and ground-breaking applications in genomic and biomedical machine learning.