Artificial Intelligence Applications and Innovations

Due to the surge of interest in online retailing, the use of credit cards has been rapidly expanded in recent years. Stealing the card details to perform online transactions, which is called fraud, has also seen more frequently. Preventive solutions and instant fraud detection methods are widely studied due to critical financial losses in many industries. In this work, a Gradient Boosting Tree (GBT) model for the real-time detection of credit card frauds on the streaming Card-Not-Present (CNP) transactions is investigated with the use of different attributes of card transactions. Numerical, hand-crafted numerical, categorical and textual attributes are combined to form a feature vector to be used as a training instance. One of the contributions of this work is to employ transaction aggregation for the categorical values and inclusion of vectors from a character level word embedding model which is trained on the merchant names of the transactions. The other contribution is introducing a new strategy for training dataset generation employing the sliding window approach in a given time frame to adapt to the changes on the trends of fraudulent transactions. In the experiments, the feature engineering strategy and the automated training set generation methodology are evaluated on the real credit card transactions.

The emerging 5G networks promises more throughput, faster, and more reliable services, but as the network complexity and dynamics increases, it becomes more difficult to troubleshoot the systems. Vendors are spending a lot of time and effort on early anomaly detection in their development cycle and majority of the time is spent on manually analyzing system logs. While main research in anomaly detection uses performance metrics, anomaly detection using functional behaviour is still lacking in depth analysis. In this paper we show how a boosted ensemble of Long Short Term Memory classifiers can detect anomalies in the 5G Radio Access Network system logs. Acquiring system logs from a live 5G network is difficult due to confidentiality issues, live network disturbance, and problems to repeat scenarios. Therefore, we perform our evaluation on logs from a 5G test bed that simulate realistic traffic in a city. Our ensemble learns the functional behaviour of an application by training on logs from normal execution time. It can then detect deviations from normal behaviour and also be retrained on false positive cases found during validation. Anomaly detection in RAN shows that our ensemble called BoostLog, outperforms a single LSTM classifier and further testing on HDFS logs confirms that BoostLog also can be used in other domains. Instead of using domain experts to manually analyse system logs, BoostLog can be used by less experienced trouble shooters to automatically detect anomalies faster and more reliable.

The most common ways of measuring Cognitive Load (CL) is using physiological sensor signals e.g., Electroencephalography (EEG), or Electrocardiogram (ECG). However, these signals are problematic in situations e.g., in dynamic moving environments where the user cannot relax with all the sensors attached to the body and it provides significant noises in the signals. This paper presents a case study using a contact-free approach for CL classification based on Heart Rate Variability (HRV) collected from ECG signal. Here, a contact-free approach i.e., a camera-based system is compared with a contact-based approach i.e., Shimmer GSR+ system in detecting CL. To classify CL, two different Machine Learning (ML) algorithms, mainly, Support Vector Machine (SVM) and k-Nearest-Neighbor (k-NN) have been applied. Based on the gathered Inter-Beat-Interval (IBI) values from both the systems, 13 different HRV features were extracted in a controlled study to determine three levels of CL i.e., S0: low CL, S1: normal CL and S2: high CL. To get the best classification accuracy with the ML algorithms, different optimizations such as kernel functions were chosen with different feature matrices both for binary and combined class classifications. According to the results, the highest average classification accuracy was achieved as 84% on the binary classification i.e. S0 vs S2 using k-NN. The highest F1 score was achieved 88% using SVM for the combined class considering S0 vs (S1 and S2) for contact-free approach i.e. the camera system. Thus, all the ML algorithms achieved a higher classification accuracy while considering the contact-free approach than contact-based approach.

In this paper, we suggest a real-time online shopper behavior prediction system which predicts the visitor’s shopping intent as soon as the website is visited. To do that, we rely on session and visitor information and we investigate naïve Bayes classifier, C4.5 decision tree and random forest. Furthermore, we use oversampling to improve the performance and the scalability of each classifier. The results show that random forest produces significantly higher accuracy and F1 Score than the compared techniques.

Smart cities emerge as highly sophisticated bionetworks, providing smart services and ground-breaking solutions. This paper relates classification with Smart City projects, particularly focusing on traffic prediction. A systematic literature review identifies the main topics and methods used, emphasizing on various Smart Cities components, such as data harvesting and data mining. It addresses the research question whether we can forecast traffic load based on past data, as well as meteorological conditions. Results have shown that various models can be developed based on weather data with varying level of success.

With the advent of social media, concepts such as forecasting and now casting became part of the public debate. Past successes include predicting election results, stock prices and forecasting events or behaviors. This work aims at using Twitter data, related to songs and artists that appeared on the top 10 of the Billboard Hot 100 charts, performing sentiment analysis on the collected tweets, to predict the charts in the future. Our goal was to investigate the relation between the number of mentions of a song and its artist, as well as the semantic orientation of the relevant posts and its performance on the subsequent chart. The problem was approached via regression analysis, which estimated the difference between the actual and predicted positions and moderated results. We also focused on forecasting chart ranges, namely the top 5, 10 and 20. Given the accuracy and F-score achieved compared to previous research, our findings are deemed satisfactory, especially in predicting the top 20.

As the requirements for automatic speech recognition are continually increasing, the demand for accuracy and efficiency is also of particular interest. In this paper, we present most of the well-known Automated Speech Recognition systems (ASR), and we benchmark three of them, namely the IBM Watson, Google, and Wit, using the WER, Hper, and Rper error metrics. The experimental results show that Google’s automatic speech recognition performs better among the three systems. We intend to extend the benchmarking both to include most of the available Automated Speech Recognition systems and increase our test data.

The amount of devices gathering and using personal data without the person’s approval is exponentially growing. The European General Data Protection Regulation (GDPR) came following the requests of individuals who felt at risk of personal privacy breaches. Consequently, privacy preservation through machine learning algorithms were designed based on cryptography, statistics, databases modeling and data mining. In this paper, we present two-levels data anonymization methods. The first level consists of anonymizing data using an unsupervised learning protocol, and the second level is anonymization by incorporating the discriminative information to test the effect of labels on the quality of the anonymized data. The results show that the proposed approaches give good results in terms of utility what preserves the trade-off between data privacy and its usefulness.

This paper introduces a novel graph-based approach to select features from multiple textual documents. The proposed solution enables the investigation of the importance of a term into a whole corpus of documents by utilizing contemporary graph theory methods, such as community detection algorithms and node centrality measures. Compared to well-tried existing solutions, evaluation results show that the proposed approach increases the accuracy of most text classifiers employed and decreases the number of features required to achieve ‘state-of-the-art’ accuracy. Well-known datasets used for the experimentations reported in this paper include 20Newsgroups, LingSpam, Amazon Reviews and Reuters.

We present a novel justification why k-means clusters should be (hyper)ball-shaped ones. We show that the clusters must be ball-shaped to attain motion-consistency. If clusters are ball-shaped, one can derive conditions under which two clusters attain the global optimum of k-means. We show further that if the gap is sufficient for perfect separation, then an incremental k-means is able to discover perfectly separated clusters. This is in conflict with the impression left by an earlier publication by Ackerman and Dasgupta. The proposed motion-transformations can be used to the new labeled data for clustering from existent ones.

finElink is a recommendation system that provides guidance to French innovative companies with regard to their financing strategy through public funding mechanisms. Analysis of financial data from former funding recipients partially feeds the recommendation system. Financial company data from a representative French population are reduced and projected onto a two-dimensional space with Uniform Manifold Approximation and Projection, a manifold learning algorithm. Former French funding recipients’ data are projected onto the two-dimensional space. Their distribution is non-uniform, with data concentrating in one region of the projection space. This region is identified using Density-based Spatial Clustering of Applications with Noise. Applicant companies which are projected within this region are labeled potential funding recipients and will be suggested the most competitive funding mechanisms.

The more complex data are, the higher the number of possibilities to extract partial information from those data. These possibilities arise by adopting different analytic approaches. The heterogeneity among these approaches and in particular the heterogeneity in results they produce are challenging for follow-up studies, including replication, validation and translational studies. Furthermore, they complicate the interpretation of findings with wide-spread relevance. Here, we take the example of statistical epistasis networks derived from genome-wide association studies with single nucleotide polymorphisms as nodes. Even though we are only dealing with a single data type, the epistasis detection problem suffers from many pitfalls, such as the wide variety of analytic tools to detect them, each highlighting different aspects of epistasis and exhibiting different properties in maintaining false positive control. To reconcile different network views to the same problem, we considered 3 network aggregation methods and discussed their performance in the context of epistasis network aggregation. We furthermore applied a latent class method as best performer to real-life data on inflammatory bowel disease (IBD) and highlighted its benefits to increase our understanding about IBD underlying genetic architectures.

While several application domains are exploiting the added-value of analytics over various datasets to obtain actionable insights and drive decision making, the public policy management domain has not yet taken advantage of the full potential of the aforementioned analytics and data models. Diverse and heterogeneous datasets are being generated from various sources, which could be utilized across the complete policies lifecycle (i.e. modelling, creation, evaluation and optimization) to realize efficient policy management. To this end, in this paper we present an overall architecture of a cloud-based environment that facilitates data retrieval and analytics, as well as policy modelling, creation and optimization. The environment enables data collection from heterogeneous sources, linking and aggregation, complemented with data cleaning and interoperability techniques in order to make the data ready for use. An innovative approach for analytics as a service is introduced and linked with a policy development toolkit, which is an integrated web-based environment to fulfil the requirements of the public policy ecosystem stakeholders.

Products with short life cycles are becoming increasingly common in many industries due to higher levels of competition, shorter product development time and increased product diversity. Accurate demand forecasting of such products is crucial as it plays an important role in driving efficient business operations and achieving a sustainable competitive advantage. Traditional forecasting methods are inappropriate for this type of products due to the highly uncertain and volatile demand and the lack of historical sales data. It is therefore critical to develop different forecasting methods to analyse the demand trend of these products. This paper proposes a new data mining approach based on the incremental k-means clustering algorithm and the RULES-6 rule induction classifier for forecasting the demand of short life cycle products. The performance of the proposed approach is evaluated using real data from one of the leading Egyptian companies in IT ecommerce and retail business, and results show that it has the capability to accurately forecast demand trends of new products with no historical sales data.

Recent attempts at Super-Resolution for medical images used deep learning techniques such as Generative Adversarial Networks (GANs) to achieve perceptually realistic single image Super-Resolution. Yet, they are constrained by their inability to generalise to different scale factors. This involves high storage and energy costs as every integer scale factor involves a separate neural network. A recent paper has proposed a novel meta-learning technique that uses a Weight Prediction Network to enable Super-Resolution on arbitrary scale factors using only a single neural network. In this paper, we propose a new network that combines that technique with SRGAN, a state-of-the-art GAN-based architecture, to achieve arbitrary scale, high fidelity Super-Resolution for medical images. By using this network to perform arbitrary scale magnifications on images from the Multimodal Brain Tumor Segmentation Challenge (BraTS) dataset, we demonstrate that it is able to outperform traditional interpolation methods by up to 20$$\%$$ on SSIM scores whilst retaining generalisability on brain MRI images. We show that performance across scales is not compromised, and that it is able to achieve competitive results with other state-of-the-art methods such as EDSR whilst being fifty times smaller than them. Combining efficiency, performance, and generalisability, this can hopefully become a new foundation for tackling Super-Resolution on medical images.

In this paper we introduce a fusion framework for image tampering localization, that moves towards overcoming the limitation of available tools by allowing a synergistic analysis and multiperspective refinement of the final forensic report. The framework is designed to combine multiple state-of-the-art techniques by exploiting their complementarities so as to produce a single refined tampering localization output map. Extensive evaluation experiments of state-of-the-art methods on diverse datasets have resulted in a modular framework design where candidate methods go through a multi-criterion selection process to become part of the framework. Currently, this includes a set of five passive tampering localization methods for splicing localization on JPEG images. Our experimental findings on two different benchmark datasets showcase that the fused output achieves high performance and advanced interpretability by managing to leverage the correctly localized outputs of individual methods, and even detecting cases that were missed by all individual methods.

Brain tumor classification is very important in medical applications to develop an effective treatment. In this paper, we use brain contrast-enhanced magnetic resonance images (CE-MRI) benchmark dataset to classify three types of brain tumor (glioma, meningioma and pituitary). Due to the small number of training dataset, our classification systems evaluate deep transfer learning for feature extraction using nine deep pre-trained convolutional Neural Networks (CNNs) architectures. The objective of this study is to increase the classification accuracy, speed the training time and avoid the overfitting. In this work, we trained our architectures involved minimal pre-processing for three different epoch number in order to study its impact on classification performance and consuming time. In addition, the paper benefits acceptable results with small number of epoch in limited time. Our interpretations confirm that transfer learning provides reliable results in the case of small dataset. The proposed system outperforms the state-of-the-art methods and achieve 98.71% classification accuracy.

This paper deals with the automated composition of music. Although music within the field of AI has been studied relatively extensively, the arena in which we operate is, to the best of our knowledge, unexplored. When it concerns computer composition, a noteworthy piece of research has involved the automated generation of the chordal notes when the underlying melody is specified. The chordal notes and beats have been generated based on models of well-known composers like Bach, Beethoven, and so on. The problem we study is the converse. We assume that the system is provided with the chords of some unknown melody. Our task is to generate a melody that flows with the given chords and which is also aesthetically and musically fitting. As far as we know, there is no research that has been reported to solve this problem and in that sense, although we have merely provided “baby steps”, our work is pioneering.

Predicting the activity of chemical compounds against cancer is a crucial task. Active chemical compounds against cancer help pharmaceutical drugs producers in the conception of anticancer medicines. Still the innate way of representing chemical compounds is by graphs, the machine learning algorithms can not handle directly the anticancer activity prediction problems. Dealing with data defined on a non-Euclidean domain gave rise to a new field of research on graphs. There has been many proposals over the years, that tried to tackle the problem of representation learning on graphs. In this work, we investigate the representation power of Node2vec for embedding learning over graphs, by comparing it to the theoretical framework Graph Isomorphism Network (GIN). We prove that GIN is a deep generalization of Node2vec. We then exert the two models Node2vec and GIN to extract regular representations from chemical compounds and make predictions about their activity against lung and ovarian cancer.

The study of Self-organizing lists deals with the problem of lowering the average-case asymptotic cost of a list data structure receiving query accesses in Non-stationary Environments (NSEs) with the so-called “locality of reference” property. The de facto schemes for Adaptive lists in such Environments are the Move To Front (MTF) and Transposition (TR) rules. However, significant drawbacks exist in the asymptotic accuracy and speed of list re-organization for the MTF and TR rules. This paper improves on these schemes using the design of an Adaptive list data structure as a hierarchical data “sub”-structure. In this framework, we employ a hierarchical Singly-Linked-Lists on Singly-Linked-Lists (SLLs-on-SLLs) design, which divides the list data structure into an outer and inner list context. The inner-list context is itself a SLLs containing sub-elements of the list, while the outer-list context contains these sublist partitions as its primitive elements. The elements belonging to a particular sublist partition are determined using reinforcement learning schemes from the theory of Learning Automata. In this paper, we show that the Transitivity Pursuit-Enhanced Object Migration Automata (TPEOMA) can be used in conjunction with the hierarchical SLLs-on-SLLs as the dependence capturing mechanism to learn the probabilistic distribution of the elements in the Environment. The idea of Transitivity builds on the Pursuit concept that injects a noise filter into the EOMA to filter divergent queries from the Environment, thereby increasing the likelihood of training the Automaton to approximate the “true” distribution of the Environment. By taking advantage of the Transitivity phenomenon based on the statistical distribution of the queried elements, we can infer “dependent” query pairs from non-accessed elements in the transitivity relation. The TPEOMA-enhanced hierarchical SLLs-on-SLLs schemes results in superior performances to the MTF and TR schemes as well as to the EOMA-enhanced hierarchical SLLs-on-SLLs schemes in NSEs. However, the results are observed to have superior performances to the PEOMA-enhanced hierarchical schemes in Environments with a Periodic non-stationary distribution but were inferior in Markovian Switching Environments.

Brain-inspired Hyperdimensional (HD) computing is an emerging technique for cognitive tasks in the field of low-power design. As an energy-efficient and fast learning computational paradigm, HD computing has shown great success in many real-world applications. However, an HD model incrementally trained on multiple tasks suffers from the negative impacts of catastrophic forgetting. The model forgets the knowledge learned from previous tasks and only focuses on the current one. To the best of our knowledge, no study has been conducted to investigate the feasibility of applying multi-task learning to HD computing. In this paper, we propose Task-Projected Hyperdimensional Computing (TP-HDC) to make the HD model simultaneously support multiple tasks by exploiting the redundant dimensionality in the hyperspace. To mitigate the interferences between different tasks, we project each task into a separate subspace for learning. Compared with the baseline method, our approach efficiently utilizes the unused capacity in the hyperspace and shows a 12.8% improvement in averaged accuracy with negligible memory overhead.

Authorship attribution attempts to identify the authors behind texts and has important applications mainly in cyber-security, digital humanities and social media analytics. An especially challenging but very realistic scenario is cross-domain attribution where texts of known authorship (training set) differ from texts of disputed authorship (test set) in topic or genre. In this paper, we modify a successful authorship verification approach based on a multi-headed neural network language model and combine it with pre-trained language models. Based on experiments on a controlled corpus covering several text genres where topic and genre is specifically controlled, we demonstrate that the proposed approach achieves very promising results. We also demonstrate the crucial effect of the normalization corpus in cross-domain attribution.

Over the last years, Indoor Localization Systems (ILS) evolved, due to the inability of Global Positioning Systems (GPS) to localize in indoor environments. A variety of studies tackle indoor localization with technologies such as Bluetooth Beacons and RFID that require costly installation, or techniques such as Google Wi-Fi/Cell DB and fingerprinting that leverage from the already existing Wi-FI and telecommunication infrastructure. Additionally, recent studies attempt to solve the same problem using Bio-Inspired techniques, such as Artificial Neural Networks (ANNs) and Deep Neural Networks (DNN). In this paper, we introduce a Multi-Objective Optimization Radiomap Modelling (MOO-RM) based ILS. The MOO-RM ILS divides the dataset into clusters using a K-Means algorithm and trains ANN models on the data of each cluster. The resulting models are fed into a Multi-Objective Evolutionary Algorithm based on Decomposition (MOEA/D), which minimizes the required storage space and the localization error, simultaneously. Our experimental studies demonstrate the superiority of the proposed approach on real datasets of Wi-Fi traces with respect to various existing techniques.

We investigated age related synaptic plasticity in thalamic reticular nucleus (TRN) as a part of visual information processing system in the brain. Simulation experiments were performed using a hierarchical spike timing neural network model in NEST simulator. The model consists of multiple layers starting with retinal photoreceptors through thalamic relay, primary visual cortex layers up to the lateral intraparietal cortex (LIP) responsible for decision making and preparation of motor response. All synaptic inter- and intra-layer connections of our model are structured according to the literature information. The present work extends the model with spike timing dependent plastic (STDP) synapses within TRN as well as from visual cortex to LIP area. Synaptic strength changes were forced by teaching signal typical for three different age groups (young, middle and elderly) determined experimentally from eye movement data collected by eye tracking device from human subjects preforming a simplified simulated visual navigation task.

With image processing, robots acquired visual perception skills; enabling them to become autonomous. Since the emergence of Artificial Intelligence (AI), sophisticated tasks such as object identification have become possible through inferencing Artificial Neural Networks (ANN). Be that as it may, Autonomous Mobile Robots (AMR) are Embedded Systems (ESs) with limited on-board resources. Thus, efficient techniques in ANN inferencing are required for real-time performance. This paper presents the process of optimizing ANNs inferencing using tensor-based optimization on embedded Graphical Processing Unit (GPU) with Computer Unified Device Architecture (CUDA) platform for parallel acceleration on ES. This research evaluates renowned network, namely, You-Only-Look-Once (YOLO), on NVIDIA Jetson TX2 System-On-Module (SOM). The findings of this paper display a significant improvement in inferencing speed in terms of Frames-Per-Second (FPS) up to 3.5 times the non-optimized inferencing speed. Furthermore, the current CUDA model and TensorRT optimization techniques are studied, comments are made on its implementation for inferencing, and improvements are proposed based on the results acquired. These findings will contribute to ES developers and industries will benefit from real-time performance inferencing for AMR automation solutions.

Everybody likes to make a good prediction, in particular, when some sort of personal investment is involved in terms of finance, energy or time. The difficulty is to make a prediction that optimises the reward obtained from the original contribution; this is even more important when investments are the core service offered by a business or pension fund generated by monthly contributions. The complexity of finance is that the human predictor may have other interests or bias than the human investor, the trust between predictor and investor will never be completely established as the investor will never know if the predictor has generated, intentionally or unintentionally, the optimum possible reward. This paper presents the Random Neural Network in recurrent configuration that makes predictions on time series data, specifically, prices. The biological model inspired by the brain structure and neural interconnections makes predictions entirely on previous data from the time series rather than predictions based on several uncorrelated inputs. The model is validated against the property, stock and Fintech market: 1) UK property prices, 2) stock markets indice prices, 3) cryptocurrency prices. Experimental results show that the proposed method makes accurate predictions on different investment portfolios.

In this paper, we present a novel method for jointly detecting and segmenting multiple objects from an untrimmed video. Unlike most existing video object segmentation methods that can only handle a trimmed video in which all video frames contain the target objects, we address a more practical and difficult problem, i.e., joint multi-object detection and segmentation from an untrimmed video where the target objects do not always appear per frame. In particular, our method consists of two modules, i.e., object decision module and object segmentation module. The object decision module is used to detect the objects and decide which target objects need to be separated out from video. As there are usually two or more target objects and they do not always appear in the whole video, we introduce the data association into object decision module to identify their correspondences among frames. The object segmentation module aims to separate the target objects identified by object decision module. In order to extensively evaluate the proposed method, we introduce a new dataset named UNVOSeg dataset, in which $$7.2\%$$ of the video frames do not contain objects. Experimental results on four datasets demonstrate that our method outperforms most of the state-of-the-art approaches.

Autonomous driving is conducted in complex scenarios, which requires to detect 3D objects in real time scenarios as well as accurately track these 3D objects in order to get such information as location, size, trajectory, velocity. MOT (Multi-Object Tracking) performance is heavily dependent on object detection. Once object detection gives false alarms or missing alarms, the multi-object tracking would be automatically influenced. In this paper, we propose a coupling system which combines 3D object detection and multi-object tracking into one framework. We use the tracked objects as a reference in 3D object detection, in order to locate objects, reduce false or missing alarms in a single frame, and weaken the impact of false and missing alarms on the tracking quality. Our method is evaluated on kitti dataset and is proved effective.

Appropriate indexing of resources is necessary for their efficient search, discovery and utilization. Relying solely on manual effort is time-consuming, costly and error prone. On the other hand, the special nature, volume and broadness of biomedical literature pose barriers for automated methods. We argue that current word embedding algorithms can be efficiently used to support the task of biomedical text classification. Both deep- and shallow network approaches are implemented and evaluated. Large datasets of biomedical citations and full texts are harvested for their metadata and used for training and testing. The ontology representation of Medical Subject Headings provides machine-readable labels and specifies the dimensionality of the problem space. These automated approaches are still far from entirely substituting human experts, yet they can be useful as a mechanism for validation and recommendation. Dataset balancing, distributed processing and training parallelization in GPUs, all play an important part regarding the effectiveness and performance of proposed methods.

There is great potential in interdisciplinary traveling platforms mingling knowledge about cultural heritage aspects, such as places with schedules providing visits or even containing augmented reality features also, along with environmental concerns to enhance personalized tourist experience and tripping avocation. For an ontological framework to support and nominate trip detours of targeted interests according to end-users, it should incorporate and unify as much heterogeneous information, deriving either from web sources or wherever there are ubiquitously available such as sensors or open databases. A plethora of qualitatively diverse data along with adequate quantities of them escalate the contingent results in terms of conferring a plurality of relevant options which can be utterly manifested through involving axioms with rule-based reasoning functionalities upon properties considered to be irrelevant to each other at first glance. Thus, managing to import predefined concepts from other ontologies, such as temporality or spatiality, and combine them with new defined concepts to tourist assets, such as points of interest, results in novel meaningful relationships never established before. Apart from the utilization of pre-existent resources and logic towards automatic detouring suggestions, a wide-spectrum modeling enables a suitable problem statement relevant to the e-Tracer framework and comprehension of the issues, providing the opportunity of statistical analysis of knowledge when adequate amounts amassed.

Modelling of knowledge and actions in AI has advanced over the years but it is still a challenging topic due to the infamous frame problem, the inadequate formalization and the lack of automation. Some problems in cyber security such as logical vulnerability, risk assessment, policy validation etc. still require formal approach. In this paper we present the foundations of a new formal framework to address these challenges. Our approach is based on three-level formalisation: ontological, logical and analytical levels. Here we are presenting the first two levels which allow to model the security policies and provide a practical solution to the frame problem by efficient utilization of parameters as side effects. Key concepts are the situations, actions, events and rules. Our framework has potential use for analysis of a wide range of transactional systems within the financial, commercial and business domains and further work will include analytical level where we can perform vulnerability analysis of the model.

Large ontologies are available as linked data, and they are used across many domains, but to process them considerable resources are required. RDF provides automation possibilities for semantic interpretation, which can lower the effort. We address the usage of RDF reasoning in large ontologies, and we test approaches for solving reasoning problems, having in mind use cases of low availability of computational resources. In our experiment, we designed and evaluated a method based on a reasoning problem of inferring Schema.org statements from cultural objects described in Wikidata. The method defines two intermediate tasks that reduce the volume of data used during the execution of the RDF reasoner, resulting in an efficient execution taking on average 10.3 ± 7.6 ms per RDF resource. The inferences obtained in the Wikidata test were analysed and found to be correct, and the computational resource requirements for reasoning were significantly reduced. Schema.org inference resulted in at least one rdf:type statement for each cultural resource, but the inference of Schema.org predicates was below expectations. Our experiment on cultural data has shown that Wikidata contains alignment statements to other ontologies used in the cultural domain, which with the application of RDF and OWL reasoning can be used to infer views of Wikidata expressed in cultural domain’s data models.

Sentiment analysis is a vigorous research area, with many application domains. In this work, aspect-based sentiment prediction is examined as a component of a larger architecture that crawls, indexes and stores documents from a wide variety of online sources, including the most popular social networks. The textual part of the collected information is processed by a hybrid bi-directional long short-term memory architecture, coupled with convolutional layers along with an attention mechanism. The extracted textual features are then combined with other characteristics, such as the number of repetitions, the type and frequency of emoji ideograms in a fully-connected, feed-forward artificial neural network that performs the final prediction task. The obtained results, especially for the negative sentiment class, which is of particular importance in certain cases, are encouraging, underlying the robustness of the proposed approach.

The main goal of this paper is to evaluate the usability of several algorithms on various sentiment-labeled datasets. The process of creating good semantic vector representations for textual data is considered a very demanding task for the research community. The first and most important step of a Natural Language Processing (NLP) system, is text preprocessing, which greatly affects the overall accuracy of the classification algorithms. In this work, two vector space models are created, and a study consisting of a variety of algorithms, is performed on them. The work is based on the IMDb dataset which contains movie reviews along with their associated labels (positive or negative). The goal is to obtain the model with the highest accuracy and the best generalization. To measure how well these models generalize in other domains, several datasets, which are further analyzed later, are used.

Automated opinion mining of consumer reviews is becoming increasingly important due to the rising influence of reviews on online retail shopping. Existing approaches to automated opinion classification rely either on sentiment lexicons or supervised machine learning. Deep neural networks perform this classification task particularly well by utilizing dense document representation in terms of word embeddings. However, this representation model does not consider the sentiment polarity or sentiment intensity of the words. To overcome this problem, we propose a novel model of deep neural network with word-sentiment associations. This model produces richer document representation that incorporates both word context and word sentiment. Specifically, our model utilizes pre-trained word embeddings and lexicon-based sentiment indicators to provide inputs to a deep feed-forward neural network. To verify the effectiveness of the proposed model, a benchmark dataset of Amazon reviews is used. Our results strongly support integrated document representation, which shows that the proposed model outperforms other existing machine learning approaches to opinion mining of consumer reviews.

In recent years, many models for predicting movie ratings have been proposed, focusing on utilizing movie reviews combined with sentiment analysis tools. In this study, we offer a different approach based on the emotionally analyzed concatenation of movie script and their respective reviews. The rationale behind this model is that if the emotional experience described by the reviewer corresponds with or diverges from the emotions expressed in the movie script, then this correlation will be reflected in the particular rating of the movie. We collected a dataset consisting of 747 movie scripts and 78.000 reviews and recreated many conventional approaches for movie rating prediction, including Vector Semantics and Sentiment Analysis techniques ran with a variety of Machine Learning algorithms, in order to more accurately evaluate the performance of our model and the validity of our hypothesis. The results indicate that our proposed combination of features achieves a notable performance, similar to conventional approaches.

MuseLearn is a platform that enhances the presentation of the exhibits of a museum with multimedia-rich content that is adapted and recommended for certain visitor profiles and playbacks on their mobile devices. The platform consists mainly of a content management system that stores and prepares multimedia material for the presentation of exhibits; a recommender system that monitors objectively the visitor’s behavior so that it can further adapt the content to their needs; and a pose estimation system that identifies an exhibit and links it to the additional content that is prepared for it. We present the systems and the initial results for a selected set of exhibits in Herakleidon Museum, a museum holding temporary exhibitions mainly about ancient Greek technology. The initial evaluation that we presented is encouraging for all systems. Thus, the plan is to use the developed systems for all museum exhibits as well as to enhance their functionality.

This work proposes an efficient Content-Based (CB) product recommendation methodology that promotes diversity. A heuristic CB approach incorporating feature weighting and Locality-Sensitive Hashing (LSH) is used, along with the TF-IDF method and functionality of tuning the importance of product features to adjust its logic to the needs of various e-commerce sites. The problem of efficiently producing recommendations, without compromising similarity, is addressed by approximating product similarities via the LSH technique. The methodology is evaluated on two sets with real e-commerce data. The evaluation of the proposed methodology shows that the produced recommendations can help customers to continue browsing a site by providing them with the necessary “next step”. Finally, it is demonstrated that the methodology incorporates recommendation diversity which can be adjusted by tuning the appropriate feature weights.