The International Classification of Diseases (ICD) serves as a standard in medical coding. Researchers in artificial intelligence, including those focused on natural language processing and machine learning, have made a significant effort to build and develop automatic ICD encoding systems and algorithms. Many algorithms have been developed to implement automatic ICD encoding, but almost all of these algorithms depended on the raw text input without taking into consideration the important medical entities in this input. In this paper, we propose an algorithm for automatically predicting ICD codes based on patient claims. Our algorithm contains several steps for finding the most relevant ICD codes. Primarily, our proposed algorithm employs medical named entity recognition (NER) to find the most important medical entities in a patient claim. For this purpose, the Medical NER model was used based on the BERT model. Next, the algorithm generates embeddings for the extracted entities using the ClinicalBERT model. To identify the most relevant ICD code, the algorithm creates embeddings for an ICD catalog, which contains various information such as chapter descriptions, long descriptions, short descriptions, and ICD codes. The embedding process is primarily based on the long descriptions, and the results are stored in a local database that contains embedding vectors and corresponding mapped ICD codes. The final step of the algorithm calculates the cosine similarity between the embedding vector generated from the patient complaint and the ICD long description vectors. The strength of this new algorithm is that it first detects the medical entities in the textual input and then predicts the most similar ICD codes. Also, our developed algorithm does not need such huge data for training. We tested the developed algorithm on a medical dataset, and the results indicate that the proposed method is highly efficient, achieving a precision rate of approximately 90%.

Social networks on the Internet have become a home that attracts all types of human thinking to exchange knowledge and ideas and share businesses. On the other hand, it has also become a source for researchers to analyze this knowledge and frame it in patterns that define types of thoughts circulating on these networks and representing the communities around them. In particular, some social networks on the Dark Web attract a special kind of thinking centered around the malicious and illegal activities disseminated on websites and marketplaces on the Dark Web. These networks involve discussions to exchange and discourse information, tips, and advice on performing such business. Studying social networks on the Dark Web is still in its infancy. In this paper, we present a methodology for analyzing the content of social networks on the Dark Web using topic modeling methods. We demonstrate the needed stages for the topic modeling process, beginning with data preprocessing and feature extraction to topic modeling algorithms. We utilize and discuss the following four topic models: LDA, CTM, PAM, and PTM. We discuss the following four topic coherence measures as evaluation metrics: UMass, UCI, CNPMI, and CV, demonstrating the selection of the best number of topics for each model according to the most coherent produced topics. Furthermore, we discuss the limitations, challenges, and future work. Our proposed approach highlights the ability to discover the latent thematic patterns in conversations and messages in the common language used in social networks on the Dark Web, constructing topics as groups of terms and their associations. This paper provides researchers with a leading methodology for analyzing thought patterns on the Dark Web.

Due to the sanctions imposed on Syria on an ongoing basis, the most recent of which is the “Caesar Act,” which has affected Syria in general and in particular the information sector, and whose effects included the ban on the use of digital certificates (SSL Certificate); it has become impossible to use locally generated digital certificates granted by the National Authority for Information Technology Services in Syria, because browsers will not support the use of these certificates specific to this country. Syria has previously obtained, through the Ministry of Communications and Technology, in cooperation with a friendly country, the authority to issue national digital certificates and a national digital certificate generator. However, after the issuance of the “Caesar Act,” which stipulates in one of its paragraphs the imposition of information sanctions on Syria, the authentication of the Syrian digital certificate issuance authority has been prevented by international authentication authorities, including (VeriSign, Thawte, Geotrust, Comodo, Entrust, DigiCert, and others). Consequently, the authority to issue Syrian digital certificates becomes unreliable because international browsers will not accept their use except after obtaining electronic permission to certify their authenticity and reliability by international authentication authorities, and the latter will not grant acceptance due to sanctions. This research presents a proposed technique for addressing the aforementioned ban problem, by building an open-source web browser that does not depend on international authentication authorities to authenticate nationally generated digital certificates, but rather it breaks the connection with those international authentication authorities and ensures correct operation of using these national certificates.

Social networks on the dark web are rich in data that provides valuable insight into the nature of the activities on the dark web and human behaviors related to these activities. It also encompasses a diversity of ideologies, interests, and thought patterns associated with illicit activities and businesses on the dark web. For this reason, social networks on the dark web constitute a powerful tool and a profuse data source for various investigative work. However, such investigations encounter considerable challenges related to the massive volumes of textual data, analyzing it effectively, and extracting knowledge from it. This knowledge can be used in various investigations and studies when representing it in ontologies as a unified and integrative data source. In this paper, we introduce a novel approach for extracting and representing knowledge hidden in dark web communities through topic modeling and ontology learning methods. We start from the conceptual design of the ontology and employ several stages of text processing and analysis to achieve the desired knowledge graph, DarkOnto. These stages include data cleaning and preprocessing, topic modeling using correlated topic model (CTM), class-topic similarity estimation, ontology construction, ontology population, and ontology evaluation, where the proposed approach achieved high results. Furthermore, we discuss the results, limitations, challenges, and future work. This paper presents a promising approach for extracting hidden valuable knowledge from dark web communities where investigating and conceptualizing criminal communities can be conducted efficiently.

The use of semantic web technologies is very limited in Arab libraries, even with the existence of important historical and literary documents, and the great digitization movement now, the documents are still trapped within the databases of the separated libraries and are not visible to search engines.The research aims to Study the applications of semantic web technologies in catalogs of digital, hybrid libraries and catalogs of public libraries, exploring the main ontologies and conceptual Data models used in the field, and the possibilities of reusing them with Arab libraries, or the need to build new ontologies, studying the necessary steps, tools, and best practices, define the difficulties and challenges facing the Arabic semantic library and searching for possible technical solutions, our experiments target the heterogeneous databases of small, medium or large libraries, considering that some Arabic library catalogs are not following a standard Data Model, to contribute to the research and development of integrating bibliographic data of Arab libraries and moving towards the Semantic web and Linked Open Data.

Abstract Ontology integration involves combining several data sources into a single and unified knowledge representation. In the context of medical texts, ontology integration plays a crucial role, as these texts include essential knowledge for clinical and research applications. However, existing methods for incorporating ontologies into medical texts have exhibited limitations in terms of comprehensiveness, flexibility, semantic accuracy, logical rigor, efficiency, and overall effectiveness. To address these shortcomings, this paper introduces an enhanced ontology integration mechanism tailored specifically for medical texts. Our proposed mechanism integrates various data sources within medical texts seamlessly by employing a combination of ontological, logical, lexical, structural, semantic, declarative, and machine-learning techniques. This mechanism leverages a common upper ontology and a set of transformation rules to align concepts and relationships across different source ontologies. Additionally, it utilizes a supervised machine learning approach to predict mappings between concepts and relationships originating from various source ontologies. To assess the effectiveness of our mechanism, extensive evaluations using diverse datasets, ontologies, and benchmark tests for ontology matching, evaluation, text processing, and application were undertaken. Through rigorous comparisons with existing approaches, the superior performance and effectiveness of our mechanism were demonstrated. In addition, a comprehensive analysis of its strengths, weaknesses, and implications for both research and real-world implementation within the field of medicine was also performed. This mechanism represents a significant advancement in ontology integration for medical texts, empowering data-driven decision-making in the field of medicine.

Conventionally, the Internet consists of three parts: Surface, Deep, and Dark Webs. In the last two decades, a massive increase in illicit activities took place on the different platforms of the Dark Web. Moreover, social networks on Dark Web implicate extremism dissemination on a wide scale. In this paper, we propose an approach to generate textual patterns from discussions on Dark Web terrorist forums employing Data Mining techniques. The discovered patterns help identify the influential members and extract critical topics. We describe our system modules that perform data preprocessing, text preprocessing with TF-IDF weighting, outlier detection, clustering evaluation, clustering, and clustering validation, implemented with the RapidMiner tool. We apply K-Means as the Clustering method with different distance metrics, evaluate the clustering process using Elbow and Silhouette methods, and validate it using Davies-Bouldin Index. Furthermore, we investigate the effects of altering the distance metrics for outlier detection on the Clustering results.

Advancements in technology have significantly influenced members’ lives, like e-Health implementation, which brings numerous benefits to the healthcare sector, including cost reduction. However, developing countries are facing challenges in adopting e-health solutions, including electronic health record (EHR) systems. This research conducts a meta-analysis, focusing on the sustainable adoption of e-Health in developing countries, using Syria as a case study. The study analyzes 20 peer-reviewed articles collected from online databases. The thematic analysis identifies four interconnected themes: social, economic, cultural, and organizational factors influencing the successful adoption of e-health in developing countries like Syria. Understanding these factors is vital for devising effective strategies to promote the sustainable integration of e-Health technologies in the healthcare systems of developing nations.

The field of Automatic Speaker Recognition (ASR) is an important and open field for researchers and scientists, especially as it has become essential in facilitating the work we do in our daily lives.Such as digital authentication and electronic transactions, and consider It as a secure environment to authenticate users' access to their accounts.Many technologies have been developed in the field of recognition but so far, there is no complete tool or method for speaker identification, the most important step in ASR is the extraction of voice features.Many methods and tools can be used to extract the speaker's vocal characteristics (voice features), which in turn will identify the user and recognize his voice spectrum through the phonetic linguistic message.In this paper, two methods will be studied, each using a different technique MFCC, which uses a logarithmic scale, and LPC, which uses a linear scale.The method used in ASR should have minimal error because it is an important authentication technology like a fingerprint, where two different people cannot have the same voice spectral range (voiceprint).

Speech recognition is one of the most important research fields nowadays because of its necessity in our daily lives and to raise the fields of security to the highest level, It's a task of speech processing, and our main scope in this paper is on speaker verification, which is to identify persons from their voices where the process depends on digitizing the sound waves into a form that allows the system to deal with it.The verification process is based on the characteristics of the speaker's voice (voice biometrics) and sends it to a further process to extract the features of that voice using the feature extraction method and using AI techniques to perform the task of identification.MFCC is used for the task of features extraction and obtains the spectrogram of a given voice signal where it represents a bank of information about the voice and sends it to the CNN model for further processing for training the model on that signal to verify if the voice belongs to a user in the system or it's a new enrollment.

From proactive detection of cyberattacks to the identification of key actors, analyzing contents of the Dark Web plays a significant role in deterring cybercrimes and understanding criminal minds. Researching in the Dark Web proved to be an essential step in fighting cybercrime, whether with a standalone investigation of the Dark Web solely or an integrated one that includes contents from the Surface Web and the Deep Web. In this review, we probe recent studies in the field of analyzing Dark Web content for Cyber Threat Intelligence (CTI), introducing a comprehensive analysis of their techniques, methods, tools, approaches, and results, and discussing their possible limitations. In this review, we demonstrate the significance of studying the contents of different platforms on the Dark Web, leading new researchers through state-of-the-art methodologies. Furthermore, we discuss the technical challenges, ethical considerations, and future directions in the domain.

The speaker identification is one of the fundamental problems in speech processing and voice modeling. The speaker identification applications include authentication in critical security systems and the accuracy of the selection. Large-scale voice recognition applications are a major challenge. Quick search in the speaker database requires fast, modern techniques and relies on artificial intelligence to achieve the desired results from the system. Many efforts are made to achieve this through the establishment of variable-based systems and the development of new methodologies for speaker identification. Speaker identification is the process of recognizing who is speaking using the characteristics extracted from the speech's waves like pitch, tone, and frequency. The speaker's models are created and saved in the system environment and used to verify the identity required by people accessing the systems, which allows access to various services that are controlled by voice, speaker identification involves two main parts: the first part is the feature extraction and the second part is the feature matching.

In this work, we introduce a novel framework based on Generative Adversarial Networks to control the pose, expression and facial features of a given face image using another face image. It can then be used for data augmentation, pose invariant face identification, face verification, and lightweight image editing. Generating new realistic face images with controllable poses, facial features, and expressions is a challenging generative learning problem due to skin tone variations, the identity preservation problem, necessity to deal with unseen large poses, and the absence of ground truth images in the training process. We make the following contributions. First, we present a network, CtrlFaceNet that can control a source face image while preserving the identity and skin tone. Second, we introduce a method for training the framework in fully self-supervised mode using a large-scale dataset of unconstrained face images. Third, we show that the style loss function can be used to preserve the skin tone of the source image. The experimental results show that our approach outperforms all other baselines. Furthermore, to the best of our knowledge, we are the first to train such a model using large-scale dataset of unconstrained face images. (C) 2020 Elsevier B.V. All rights reserved.

As the data on the online social networks is getting larger, it is important to build personalized recommendation systems that recommend suitable content to users, there has been much research in this field that uses conceptual representations of text to match user models with best content. This article presents a novel method to build a user model that depends on conceptual representation of text by using ConceptNet concepts that exceed the named entities to include the common-sense meaning of words and phrases. The model includes the contextual information of concepts as well, the authors also show a novel method to exploit the semantic relations of the knowledge base to extend user models, the experiment shows that the proposed model and associated recommendation algorithms outperform all previous methods as a detailed comparison shows in this article.

In the last two decades, illicit activities have dramatically increased on the Dark Web.Every year, Dark Web witnesses establishing new markets, in which administrators, vendors, and consumers aim to illegal acquisition and consumption.On the other hand, this rapid growth makes it quite difficult for law and security agencies to detect and investigate all those activities with manual analyses.In this paper, we introduce our approach of utilizing data mining techniques to produce useful patterns from a dark web market contents.We start from a brief description of the methodology on which the research stands, then we present the system modules that perform three basic missions: crawling and extracting the entire market data, data pre-processing, and data mining.The data mining methods include generating Association Rules from products' titles, and from the generated rules, we infer conceptual compositions vendors use when promoting their products.Clustering is the second mining aspect, where the system clusters vendors and products.From the generated clusters, we discuss the common characteristics among clustered objects, find the Top Vendors, and analyze products promoted by the latter, in addition to the most viewed and sold items on the market.Overall, this approach helps in placing a dark website under investigation.

Internet and network technologies have evolved dramatically in the last two decades, with rising users' demands to preserve their identities and privacy.Researchers have developed approaches to achieve users' demands, where the biggest part of the internet has formed, the Deep Web.However, as the Deep Web provides the resort for many benign users who desire to preserve their privacy, it also became the perfect floor for hosting illicit activities, which generated the Dark Web.This leads to the necessity of finding automated solutions to support law and security agencies in collecting information from the Dark Web to disclose such activities.In this paper, we illustrate the concepts needed for the development of a crawler that collects information from a dark website.We start from discussing the three layers of the Internet, the characteristics of the hidden and private networks, and the technical features of Tor network.We also addressed the challenges facing the dark web crawler.Finally, we presented our experimental system that fetches data from a dark market.This approach helps in putting a single dark website under investigation, and can be a seed for future research and development.

In unconstrained facial images, large visual variations concerning pose, scale, the presence of occlusions, expressions and lighting usually cause difficulties in discriminating faces from the background accurately. As a result, some non-face regions are recognized as faces (false positive) and that influences the effectiveness of face detection algorithms which is characterized by low false positive (FP) rate, high detection rate and high speed of processing. In order to reduce these non-face regions, they are considered as anomalies and then try to detect them. In this paper, we propose an anomaly detection method using reconstruction error from variational autoencoder (VAE), which is a generative machine learning model. We train VAE to learn reconstructing faces that are close to its original input faces using FDDB dataset, then the difference between the original input face and the reconstructed output is measured to obtain the reconstruction error which can be used as an anomaly score. Consequently, the regions resulting from faces detection algorithm with high reconstruction error are defined as anomalies or false positives.

As E-learning Tools and techniques are becoming more common and compelling, many researches have emerged lately that aims at making it more flexible and applicable. Besides, the content is getting very large nowadays, so that it is very important to develop a more accurate and robust search techniques that help users find the best learning materials that exists all along the web specially on social learning websites. In this paper we propose a new method to collect, index and retrieve learning materials, a collection algorithm is presented that can bring together content from various sources. We present a semantic indexing method that aims at weighting words of the document based on both DBpedia and Wordnet Ontologies, which proves more accurate results according to the analysis and comparison that are shown in this paper.