International Joint Conference SOCO’18-CISIS’18-ICEUTE’18

In this article is presented a review of the state of the art on Financial Technology (Fintech) for the design of a novel recommender system. A social computing platform is proposed, based on Virtual Organizations (VOs), that allows to improve user experience in actions that is associated with the process of investment recommendation. The work presents agents functionalities and an algorithm that will improve the accuracy of the Recommender_agent which is in charge of the Case-based reasoning (CBR) system. The data that will be collected and will feed the CBR corresponds to user’s characteristics, the asset classes, profitability, interest rate, history stock market information and financial news published in the media.

The creation of a suitable charging infrastructure for electric vehicles (EV) is one of the main challenges to increase the adoption of this new vehicle technologies. In this article, we present a Multi-Agent System (MAS) that performs an analysis of a set of possible configurations for the location of EV charging stations in a city. To estimate the best configurations, the proposed MAS considers data from heterogeneous sources such as traffic, social networks, population, etc. Based on this information, the agents are able to analyze a large set of configurations using a genetic algorithm that optimizes the configurations taking into account a utility function.

The large amounts of information that social networks contain, makes it necessary for them to provide guides and aids that improve users’ experience in the system. In addition to search and filtering tools, users should be presented with the content they wish to obtain before they take any action to find it. To be able to recommend content to users, it is necessary to analyse their profiles and determine what type of content they want to view. The present work is focused on an employability oriented social network for which a job offer recommender system is proposed, following the model of a multi-agent system. The recommendation system has a hybrid approach, consisting of a CBR system and an argumentation framework. The CBR system is capable of deciding, on the basis of a series of metrics and similar cases stored in the system, whether a job offer is likely to be recommended to a user. Besides, the argumentation framework extends the system with an argumentation CBR, through which old and similar cases can be obtained from the CBR system. Finally, based on the different solutions proposed by the agents and the experience gained from past cases, a process of discussion among agents is established. Here, a debate is held in which a final decision is reached, giving the best recommendation to the proposed problem.

The evaluation and diagnosis of retina pathologies are usually made by the analysis of different image modalities that allows to explore its structure. The most popular retina image method is the retinography, a technique to show the retina and other structures in the fundus of the eye. This paper deals with an important stage of the retina image processing for a diagnosis tool which aims to show the blood vessel structure. Our proposal is based on a deep convolutional neural network, that avoids any preprocessing stage such as gray scale conversion, histogram equalization, and other image transformations that determine the final result. Thus, we obtain the blood vessel segmentation directly from the original RGB color retinography image. The results obtained with our method are comparable to the state-of-the art methods but using a smaller network with less memory and computation requirements. Our approach has been assessed using the DRIVE database.

The automated analysis of medical imaging, especially brain imaging, is a challenging high-dimensional task. Computer Aided Diagnosis (CAD) tools often require the images to be spatially normalized and then perform feature extraction to be able to avoid the small sample size problem. However, the spatial normalization often introduces artefacts, especially in functional imaging. Furthermore, variance-based decomposition techniques like PCA, which are extensively used in CAD tools, often perform poorly in highly-unbalanced dataset. To overcome these two problems, we propose a deep Convolutional Autoencoder (CAE) architecture that performs image decomposition -or encoding- in images that were not spatially normalized. A CAD system that used CAE for feature extraction and a Support Vector Machine Classifier (SVC) for classification was tested on a strongly imbalanced (5.69/1) Parkinson’s Disease (PD) neuroimaging dataset from the Parkinson’s Progression Markers Initiative (PPMI), achieving more than 93% accuracy in detecting PD with DaTSCAN imaging, and a area under the ROC curve higher than 0.96. This system paves the way for new deep learning decompositions that bypass the common spatial normalization step and are able to extract relevant information in highly-imbalanced datasets.

In this paper we describe a system aimed at obtaining several geometric measures from seamless steel tubes representative of their manufacturing quality. Traditionally, these measurements are taken manually using calipers and similar tools, which are error-prone and have low reproducibility. We introduce a system based on 3D scanning and computational geometry. Scanning adds surface noise and may leave small undetectable gaps depending on surface properties, and the tube itself may have small defects not representative of the overall tube shape. Also, processing high resolution scans containing millions of points can be too time consuming. Thus, special algorithms are needed to make the problem tractable and produce representative measures in a reasonable time.

Power utilities experience large losses of electricity in distribution from power plants to the end consumer. There are two types of losses: technical and non-technical. Among non-technical losses is a very prominent one: electrical fraud. In this paper we propose a new system to detect fraud. A blockchain is used to store the data collected by the WSN that monitors the power distribution grid. Using data stored in the blockchain, it is constructed directed directed acyclic graph (DAG) with non-technical losses and applied the clustering algorithm created to detect fraud. The main advantage of blockchain to our model is that every time the blockchain grows the stored data is more secure. Therefore, power utilities can perform an inspection in blockchain data stored.

Performance measures are used in Machine Learning to assess the behaviour of classifiers. Many measures have been defined on the literature. In this work we focus on Confusion Entropy (CEN), a measure based in Shannon’s Entropy. We introduce a modification of this measure that overcomes its disadvantages in the binary case that disables it as a suitable measure to compare classifiers. We compare this modification with CEN and other measures, presenting analytical results in some particularly interesting cases, as well as some heuristic computational experimentation.

It is clear that social networks have come to stay. In recent years they have become the media par excellence. The users who participate in them help to spread information quickly and easily so that everyone can benefit. Users are becoming more inclined to voice their opinions, networks are willing to listen and technology has an enormous outreach. A priori, something that seems a great advantage can become a big problem when the news spread violates an individual’s right to honour. This paper proposes a tool that detects and collects information from users who publish or disseminate offensive information to an individual. It establishes parameters that determine the level of damage certain individuals can make on social media and makes a ranking that is based on their characteristics and publications. This proposal is an example of the infinite possibilities that automatic data collection and processing provide us with. Without these technologies it would have been impossible to protect the rights of individuals on social networks, due to the large number of users.

In this work, we propose a novel imaging preprocessing step based on the use of the gradient magnitude for medical DaTSCAN SPECT images. As Parkinson’s Disease (PD) is characterized by a marked reduction of intensity at striatum area, measuring intensities in this region is considered as a good marker for this neurological disorder. To extend this idea, we have been studying how quick these values decrease. A simple way to do this was using the gradient of each image. Applying Machine Learning algorithms, we have classified the gradient images and obtained an accuracy improvement of almost 2%. These results prove that the gradient magnitude is even a better marker for PD diagnosis and opens the door to new future investigations about this pathology.

Although, imbalanced data analysis gained significant attention in the past years, it still remains an underdeveloped area of research posing many difficulties due to the difference in the number of objects in the examined classes, rendering traditional, accuracy driven machine learning methods useless. With many modern real-life applications being examples of imbalanced data classification i.e. fraud detection, medical diagnosis, oil-spills detection in satellite images or network anomaly detection, it is crucial to develop new algorithms suitable to use in such situations. One of the approaches to deal with the disproportion between the instances of objects in classes are either over- or undersampling techniques. In this paper, we propose a modification of an existing RBO algorithm. Due to the additional constraint the modified algorithm eliminates instances which may be problematic to classify. Additionally, a recursion mechanism was added in order to make the search of synthetic points more robust. The results obtained from computer experiments carried out on the benchmark datasets prove that the presented algorithm is applicable.

Accurate solar energy prediction is required for the integration of solar power into the electricity grid, to ensure reliable electricity supply, while reducing pollution. In this paper we propose a new approach based on deep learning for the task of solar photovoltaic power forecasting for the next day. We firstly evaluate the performance of the proposed algorithm using Australian solar photovoltaic data for two years. Next, we compare its performance with two other advanced methods for forecasting recently published in the literature. In particular, a forecasting algorithm based on similarity of sequences of patterns and a neural network as a reference method for solar power forecasting. Finally, the suitability of all methods to deal with big data time series is analyzed by means of a scalability study, showing the deep learning promising results for accurate solar power forecasting.

In peer-to-peer (P2P) lending, it is important to predict default of borrowers because the lenders would suffer financial loss if the borrower fails to pay money. The huge lending transaction data generated online helps to predict repayment of the borrowers, but there are limitations in extracting features based on the complex information. Convolutional neural networks (CNN) can automatically extract useful features from large P2P lending data. However, as deep CNN becomes more complex and deeper, the information about input vanishes and overfitting occurs. In this paper, we propose a deep dense convolutional networks (DenseNet) for default prediction in P2P social lending to automatically extract features and improve the performance. DenseNet ensures the flow of loan information through dense connectivity and automatically extracts discriminative features with convolution and pooling operations. We capture the complex features of lending data and reuse loan information to predict the repayment of the borrower. Experimental results show that the proposed method automatically extracts useful features from Lending Club data, avoids overfitting, and is effective in default prediction. In comparison with deep CNN and other machine learning methods, the proposed method has achieved the highest performance with 79.6%. We demonstrate the usefulness of the proposed method as the 5-fold cross-validation to evaluate the performance.

Mobile marketing is an expanding industry due to the growth of mobile devices (e.g., tablets, smartphones). In this paper, we explore a categorical approach to cluster publishers of a mobile performance market, in which payouts are only issued when there is a conversion (e.g., a sale). As a case study, we analyze recent and real-world data from a global mobile marketing company. Several experiments were held, considering a first internal evaluation stage, using training data, clustering quality metrics and computational effort. In the second stage, the best method, COBWEB algorithm, was analyzed using an external evaluation based on business metrics, computed over test data, and that allowed an identification of interesting clusters.

Modernly, main Internet traffic is by the means of wireless network, due to the high mobility requirements and surging number of portable electronic devices users. Rapid growth of demand for uninterrupted access to vast amounts of information that are available in Internet forces network operators to focus on network reliability and to anticipate potential events such as network overload, that could pose threat for sustained delivery of data. In this paper Author investigated efficiency of WiFi open network in building located at main campus of Wrocław University of Science and Technology (WUST). The database analyzed in this paper consist data from two monthly periods, namely May of 2014 and 2015. The idea of research was to create spatial model prediction of WiFi network efficiency. Models of prediction contains two important parameters of WiFi network: number of users and load channel utilization. Spatial (3D) predictions for two database were made with using geostatistical co-simultaion method Turning Bands. Obtained results were compared with each other and conclusions with future research directions to WiFi network efficiency predictions were drawn.

This study presents the application of neural models to a real-life problem in order to study the energy and water consumptions of an automotive multinational company for resources saving and environment protection. The aim is to visually and naturally analyse different consumptions data for a whole year, month by month, from factories and locations worldwide where different kinds of products are produced. The data are studied in order to see whether the geographical location, the month of the year or the technology used in each factory are relevant in terms of consumptions and then take actions for a greener production. The consumptions dataset is analysed using different neural projection models: Principal Component Analysis and Cooperative Maximum-Likelihood Hebbian Learning. This unsupervised dimensionality reduction techniques have been applied, and subsequent interesting conclusions are obtained.

The manufacturing industry is an ever-changing environment with companies facing increasing external and internal challenges, such as economic crises, technological development and global competition. These challenges create the need for companies to constantly adapt as the environment around them changes. As such, companies are adopting a more proactive approach to manufacturing rather than the usual reactive process, by taking advantage of the ongoing move towards automation and system interconnectivity in the context of Industry 4.0.In this work, we propose an agent-based architecture that presents a solution to project scheduling problems, with operation dependent setup time that is resource, material and human-resource constrained.

The increasingly complex tasks require an enormous effort in path planning within dynamic environments. This paper presents a efficient method for detecting collisions between a robot and its environment in order to prevent dangerous maneuvers. Our methods is based upon the transformation of each robot link and the environment in a set of bounding boxes. The aim of this kind of prismatic approximation is to detect a collision between objects in the workspace by testing collision between boxes from different objects. The computational cost of this approach has been tested in simulations, thus we have set up our environment with a HP20D robot and an obstacle, both represented by their corresponding chain of bounding boxes. The experiment implies to move the robot from an initial position, on the right of the obstacle, to a final position, on the left side of the obstacle, along a straight-line trajectory. The probe enabled us to check the correct behavior of a collision detector in a real situation.

Prediction is the way to optimize the maintenance task by determining the correct moment to interview, repair or replace equipment which the most difficult decision for companies in Industry 4.0 environment. This research present MASPI. I is a multiagent system based on advantages of virtual organization. The goal of MASPI is to be a reference model for making predictions about data captured by sensors installed in equipment or industrial machines. The capability of MASPI is evaluated by applying it to SCANIA trucks dataset, using machine learnings algorithms to obtain the prediction and compare their accuracy.

Evolutionary Structural Optimization (ESO) seeks to mimic the form in which nature designs shapes. This paper focuses on shape carving triggered by environmental stimuli. In this realm, existing algorithms delete under - stressed parts of a basic shape, until a reasonably efficient (under some criterion) shape emerges. In the present article, we state a generalization of such approaches in two forms: (1) We use a formalism that enables stimuli from different sources, in addition to stress ones (e.g. kinematic constraints, friction, abrasion). (2) We use metagraphs built on the Finite Element constraint graphs to eliminate the dependency of the evolution on the particular neighborhood chosen to be deleted in a given iteration. The proposed methodology emulates 2D landmark cases of ESO. Future work addresses the implementation of such stimuli type, the integration of our algorithm with evolutionary based techniques and the extension of the method to 3D shapes.

Imbalanced datasets constitute a challenge in medical-image processing and machine learning in general. When the available training data is highly imbalanced, the risk for a classifier to find the trivial solution increases dramatically. To control the risk, an estimate on the prior class probabilities is usually required. In some medical datasets, such as breast cancer imaging techniques, estimates on the priors are intractable. Here we propose a solution to the imbalanced support vector classification problem when prior estimations are absent based on a case-dependent transformation on the decision function.

Synaesthesia is a neurological phenomenon in which stimulation of one sense triggers an involuntary experience in a secondary sense. This paper draws on this phenomenon to generate a system that creates music using colors as a synaesthete input. The system optimizes a function based on the Tonal Interval Space by applying the Particle Swarm Optimization algorithm to automatically generate a tonal melody. The final result is a musical line that is evaluated in terms of musical quality.

In the paper we consider the problem of scheduling of technological operations implementation in productions cells in a company which produce steel structures of car seats. We propose an optimization algorithm based on the Branch and Bound method which determines minimal number of team members which operating production cells maintaining the maximum efficiency of the cells. The usefulness of the algorithm for practical purposes has been verified on real data.

In the paper, rationalization of production order execution in a large manufacturing company is suggested. Till now, the decision-making process was based on the human factor, which resulted in irregular utilization of manufacturing resources. The presented work was aimed at developing new order selection system. That would make it possible to utilize the admitted resources possibly best and thus to meet the deadlines and to adapt to specific production requirements. In the work, the greedy and Tabu Search algorithms were used. As a basis for the research employing, historical data were accepted. Each order were given a priority, production time, profit and penalty for failure. Additionally, the machine failure risk was calculated based on empirical measurements. Simulations of several subsequent working weeks were performed in order to analyse the results obtained thanks to the suggested methods and to compare them with the results presently reached by the company.

In manufacturing systems layout of machines has a significant impact on production time and cost. The most important reason for this is that machine layout impact on transportation time and cost. When designing layout, the aisles structure has an effect on transportation. The aisles are paths that transporters go through them to move the materials between machines. The capacity of the aisles is not infinitive and there is a limitation for the number of transporters that can pass an aisle at the same time. This causes transporters wait for the aisle to be empty or even transporters crossing. Therefore, when optimizing the layout of machines, the aisles structure and their capacity must be investigated. This paper proposes an approach for layout design in manufacturing systems taking into account capacitated aisles structure. First, the aisle structure is determined. Then, the machines are assigned in the possible areas. A simulation optimization approach is proposed to solve the problem. This helps us to avoid unrealistic assumptions and consider realistic conditions such as stochastic characteristics of the manufacturing system, random process time and random breakdown. A genetic algorithm is used to search for the best position of machines. Finally, a numerical example is included to illustrate the proposed approach.

Augmented Reality (AR) has evolved over the past years, but before it is widely adopted and used in manufacturing industry, it has to overcome a number of technological challenges. Although new advancements in tracking and display technology have been a priority in recent research works, the use of accurate registration methods is not fundamental for users to understand the intent of the augmentation. Moreover, interactive visualization of contextual data in augmented spaces did not receive enough attention from the research community. In this paper, we investigate the creation of AR workspaces focused on interaction and visualization modes rather than on the registration accuracy, and how to provide more effective means to support assembly tasks in hybrid human-machine manufacturing lines. In particular, we focus on short-lived assembly tasks, i.e. manufacturing of limited batches of customized products, which do not yield significant returns considering the effort necessary to adapt AR systems and the production time frame.

Deflectometric techniques provide abundant information useful for aesthetic defect inspection in specular and glossy/shinny surfaces. A series of light patterns is observed indirectly through their reflection on the surface under inspection, and different geometrical or texture information about the surface can be extracted. In this paper, we present a deep learning based approach for the automated defect identification in deflectometric recordings. The proposed learning framework automatically learns features used for classification. Although the method is in an early stage of development, the experiments with industrial parts show promising results, and a very direct application if compared to hand-crafted feature definition approaches.

In this work an adaptive neuro-control is proposed to cope with some external disturbances that can affect unmanned aerial vehicles (UAV) dynamics, specifically: the variation of the system mass during logistic tasks and the influence of the wind. An intelligent control strategy based on a feedforward neural networks is applied. In particular, a variant of the generalized learning algorithm has been used. Simulation results show how the on-line learning increases the robustness of the controller, reducing the effects of the changes in mass and the effects of wind on the UAV stabilization, thus improving the system response. It has been compared with a PID controller obtaining better results.

Positioning accuracy of servo drive systems is very important for tasks that require precision. From the control point of view, servo drive systems are complex due to their non-linear time-varying dynamics. Most of the control strategies applied to these systems either introduce undesirable chattering in the response or suppress it at the cost of producing large tracking errors. In this work, an equivalent sliding mode control based on fuzzy logic is applied to a servo system. The fuzzy membership functions of the switching function are optimized in order to improve the control robustness and to obtain accurate tracking at the same time. Simulation results show that this soft computing control proposal can effectively eliminate the chattering and reduce the tracking error for servo drive systems. It has been compared to conventional sliding mode control and sliding mode control with boundary layer with encouraging results.

When designing controllers for complex systems, it is not only necessary to stabilize the system but to improve the robustness in order to get a better response. Some indexes allow to measure this robustness of the system response, such as the gain and phase margins. In this paper a computational tool that implements a Multi-Objective Genetic Algorithm (MOGA) is designed and applied to optimize the robustness of different controllers. So, it is possible to analyse how the variation of the controller parameters influences the robustness of the system. The tool is applied to the optimization of a Linear Quadratic (LQ) and Eigenvalues assignment (EA) controllers for a MIMO autonomous vehicle, a helicopter, with satisfactory results.

In this work, a first approach of using the Particle Swarm Optimization (PSO) as a method for optimizing an Ensemble Model built with Extreme Learning Machines is presented. The paper focuses on the obtaining of the parameters of a weighted averaging method for a Ensemble Model, using Extreme Learning Machines as models. The main contribution of this document is the use of the heuristic algorithm PSO for searching optimum parameters of the weighted averaging method. The experiments show that PSO is suitable for computing the parameters of the ensemble, obtaining an average improvement of 68% of the error comparing with an individual model. Also other comparisons have been made with basic combining methods of Ensemble Model fulfilling the expectations.

This study focuses on the performance of a fall detection method using data coming from real falls performed by relatively young people and the application of this technique in the case of an elder person. Although the vast majority of studies concerning fall detection place the sensory on the waist, in this research the wearable device must be placed on the wrist because it’s usability. A first pre-processing stage is carried out as stated in [1, 17]; this stage detects the most relevant points to label. This study analyzes the suitability of different models in solving this classification problem: a feed-forward Neural Network and a rule based system generated with the C5.0 algorithm. A discussion about the results and the deployment issues is included.

Applications based on inverted pendulum principle, have proliferated in recent years, surveillance system such as two-wheeled vehicles or manipulators attached to robotics legs in humanoids robots are some examples. Even though the control of inverted pendulum has been deeply studied in the academic community for decades, there are few implementations with a Fuzzy Takagi Sugeno model for overlapped membership functions joint to Linear Quadratic Regulator, because just in [11], this kind of generalized T-S modeling was addressed. This paper presents the implementation of a nonlinear Fuzzy Takagi-Sugeno control with overlapped membership functions applied to a mobile inverted pendulum mechanism and its comparison against LQR in terms of state space behavior and robustness.

The forecasting of the freight transportation provides a helpful information in the management of ports environment and can be used as a decision-making tool. This work addresses the forecasting of ro-ro (roll-on roll-off) freight flow in a port using a two-stage approach by an ensemble of the best Support Vector Regression (SVR) models. The time series used for forecasting is daily ro-ro freight in the port of Algeciras during the period from 2000 to 2007. Additionally, the time series was preprocessed through an exponential smoothing in order to improve the performance. The experiment results show that the proposed approach is a promising tool in freight forecasting.

We provide a view of current research issues in Robotic Grasping and Bin Picking focused on the perception aspects of the problem, mainly related to computer vision algorithms. After recalling the evolution of the topics in the last decades, we focus on the modern use of Deep Learning Algorithms. Two main trends are followed in the approaches to innovative grasping techniques. First, Convolutional Neural Networks are used for grasping perceptual aspects. We discuss the different degrees of success of several published approaches. Second, Deep Reinforcement Learning is being extensively tested in order to develop integrated eye-hand coordination systems not requiring delicate calibration. We provide also a discussion of possible future lines of research.

The analysis of bilateral distance between home and host countries is a key issue in the internationalization strategy of companies. As a multi-faceted concept, distance encompasses multiple dimensions, with psychic distance being one of the most critical ones for the overseas investments of firms. Among all the psychic distance stimuli that have been proposed until now, the present paper focuses on Industrial Development Distance (IDD). Together with data from both the countries and the companies, IDD is analysed by means of neural projection models based on unsupervised learning, to gain deep knowledge about the internationalization strategy of Spanish large companies. Informative projections are obtained from a real-life dataset, leading to useful conclusions and the identification of those destinations attracting large flows of investment but with a particular idiosyncrasy.

Transportation is the economic activity that is the most tightly coupled with the other ones. As a result, knowledge about transportation in general, and market demand in particular, is key for an economic analyisis of a sector. In present paper, the official data about the industrial sector, coming from the Ministry of Public Works and Transport in Spain, is analysed. In order to do that, k-means clustering technique is applied to find groupings or patterns in the dataset that contains data from a whole year (2015). Samples allocation to clusters and silhouette values are used to characterize the demand of the industrial transportation. Useful insights into the analysed sector are obtained by means of the clustering technique, that has been applied with 4 different criteria.

The first step for an optimum energy consumption reducing planning supposes the accurate estimation of the primary sources. For this purpose, Light Detection and Ranging (LiDAR) remote sensing technique is being widely applied because its ability to collect huge amounts of data with good accuracy. This study focuses on the application of this technology to the improvement of the assessment of wind, solar and biomass energies. In the case of the biomass, a proof of concept of the estimation for the Pinus Radiata specie in the Arratia-Nervión region (Spain) has been explained. Due to the promising results obtained with this technique, LiDAR has stand out as a powerful and versatile tool for energy consumption reduction in the industrial sector.

The Wasman (Waste Management as Policy Tools for Corporate Governance) project is a program comprising six southern European countries that share best practices to achieve collective and intelligent decisions regarding waste management. The partnership among the actors is envisioned as a social structure in a socially distributed cognitive system that is the product of action and the conditional element of future action. The authors call this a system of “intelligent cognitive waste management.”

Systems optimization is one of the great challenges to improve the industry plants performance. From an economical point of view, a proper optimization means, among others, energy, material and maintenance savings. Furthermore, the quality of the final product is improved. So fault detection techniques development plays a very important role to achieve the system optimization. Under this topic, the present research shows the developed work over a real common system, the level control. A new proposal based on unsupervised techniques were used to detect the system malfunction states, taking into account a dataset collected during the right operation. The proposal is validated with ad-hoc created faults for the different system operation points. The performance is very satisfactory in general terms.

Inventory control is the science-based art of controlling the amount of inventory (or stock) held, in various forms. Inventory control techniques are very important components and the most organizations can substantially reduce their costs associated with the flow of materials. This paper presents biological swarm intelligence in general, and in particularly two models: particle swarm optimization and firefly algorithm for modelling on inventory control in production system. The aim of this research is to create models to minimize production cost according to price of items and inventory keeping cost.

This work proposes a model of movement detection in patients with hip surgery rehabilitation. Using the Microsoft Xbox One Kinect motion capture device, information is acquired from 25 body points -with their respective coordinate axes- of patients while doing rehabilitation exercises. Bayesian networks and sUpervised Classification System (UCS) techniques have been jointly applied to identify correct and incorrect movements. The proposed system generates a multivalent logical model, which allows the simultaneous representation of the exercises performed by patients with good precision. It can be a helpful tool to guide rehabilitation.

Biometric information extracted from electroencephalogram (EEG) signals is being used increasingly in person identification systems thanks to several advantages, compared to traditional ones such as fingerprint, face and voice. However, one of the major challenges is that a huge amount of EEG data needs to be processed, transmitted and stored. The use of EEG compression is therefore becoming necessary. Although the lossy compression technique gives a higher Compression Ratio (CR) than lossless ones, they introduce the loss of information in recovered signals, which may affect to the performance of EEG-based person identification systems. In this paper, we investigate the impact of lossy compression on EEG data used in EEG-based person identification systems. Experimental results demonstrate that in the best case, CR could achieve up to 70 with minimal loss of person identification performance, and using EEG lossy compression is feasible compared to using lossless one.

With the ever-increasing capabilities of modern hardware and breakthroughs in the DNA biometrics field, we are presenting a new, scalable and innovative method to accelerate the DNA analysis process used in criminal investigations, by building an improved methodology for using large-scale GPU-based automata for performing high-throughput pattern-matching. Our approach focuses on all important stages of preparing for the pattern-matching process, tackling with all major steps, from creation, to preprocessing, to the runtime performance. Finally, we experiment using real-world DNA sequences and apply the process to the human DNA genome, for an evaluation of our implementation.

The security of transportation is nowadays a challenge. The Intelligent Transport Systems (ITS) are included to have the newest finding over transportation features. The current work propose a new problem inspired by ITS over the Traveling Salesman Problem (TSP) highlighting the security constraints. The problem it is called the Secure Intelligent Transport Systems within the Traveling Salesman Problem (SITS-TSP). The optimization problem should satisfy the security constraints. The mathematical model is also included.

Forest fires may cause considerable damages both in ecosystems and lives. This proposal describes the application of Internet of Things and wireless sensor networks jointly with multi-hop routing through a real time and dynamic monitoring system for forest fire prevention. It is based on gathering and analyzing information related to meteorological conditions, concentrations of polluting gases and oxygen level around particular interesting forest areas. Unusual measurements of these environmental variables may help to prevent wildfire incidents and make their detection more efficient. A forest fire risk controller based on fuzzy logic has been implemented in order to activate environmental risk alerts through a Web service and a mobile application. For this purpose, security mechanisms have been proposed for ensuring integrity and confidentiality in the transmission of measured environmental information. Lamport’s signature and a block cipher algorithm are used to achieve this objective.

This paper deals with the Job shop Scheduling problem with Time Lags (JSTL). JSTL is an extension of the job shop scheduling problem, where minimum and maximum time lags are introduced between successive operations of the same job. We propose a combination between Biogeography-Based Optimization (BBO) algorithm and Greedy heuristic for solving the JSTL problem with makespan minimization. Biogeography-Based optimization is an evolutionary algorithm which is inspired by the migration of species between habitats. BBO has successfully solved optimization problems in many different domains and has reached a relatively mature state using two main steps: migration and mutation. Good performances of the proposed combination between Greedy and BBO algorithms are shown through different comparisons on benchmarks of Fisher and Thompson, Lawrence and Carlier for JSTL problem.

Plagiarism among university students is an important issue that affects their preparation and undermines the universities’ efforts to prepare skilled graduates. Universities try to fight-back this problem with strict ethics policies, but they require the proper plagiarism detection tools, at affordable costs, to implement these policies. In this paper, we present PlagZap, a cost-efficient, high-volume, and high-speed plagiarism detection system built using open-source software and designed to be used on textual student assignments (essays, theses, homework). We discuss the advantages of this design with respect to speed, precision and costs.

Logistics services involve a wide range of transport operations between distributors and clients. Currently, the large number of intermediaries are a challenge for this sector, as it makes all the processes more complicated. In this paper we propose a system that uses smart contracts to remove intermediaries and speed up logistics activities. In addition, a multi-agent system is used to coordinate entire logistics services, smart contracts and compliance with their terms. Our new model combines smart contracts and a multi-agent system to improve the current logistics system by increasing organization, security and significantly improving distribution times.

Learning a programming language requires a great deal of effort in both the theoretical and practical domains. As far as theory is concerned, a knowledge of the methods, concepts, attributes that are characteristic of the language as well an understanding of the its specific structures and peculiarities is required. On the other hand, mastering the theoretical concepts is not enough as it is necessary to be able to apply them optimally, efficiently and effectively. To adapt the teaching to those aspects that require the most attention, the weaknesses shown by the students must be identified. An exhaustive analysis of their performance – which should go beyond a mere numerical assessment – is required to focus the teaching efforts on those areas where needs are greater. Consequently, to assess the theoretical knowledge a statistical analysis from the results of the theoretical test conducted will be shown (multiple-choice type test) where the analysis is not confined to the number of wrong answers but looks at where they occur and in what percentage. As far as the practical part, a rubric has been designed to exhaustively correct the assignments, which also allows for the introduction of such remarks as are deemed necessary regarding all points of interest.

In this article, a workshop about basic artificial intelligence (AI) concepts for high school students is presented, where basic concepts of these algorithms are explained and practised using simple software tools. The aim of this activity is to proportionate a clear idea about the simple rules of AI and enhance the social trust of science since early ages.

The jump from Secondary Education to University is noticed by students once they have seen the bad obtained results after the first ordinary exam session they deal with. Many factors such as group integration problems, lack of motivation and/or required study methodology turn the jump a difficult wall to overcome. The aim of this work is start a discussion of some ideas for redesign a continuous assessment in Mathematics for Chemistry and Geology Bachelor degrees so in the near future we have a defined work lines to submit and implement an educational innovation project.

Strength of materials and elasticity are included within the core subjects in the School of Engineering. Modeling and simulation are fundamental in early stages of engineering design and testing. For this reason, we propose the use of a Java application named BEMAPEC in combination with hand made problems to improve the motivation of the engineering students. BEMAPEC applies the Boundary Element Method (BEM) to solve thermoelastic problems of one 3D solid, among other functionalities. Thus, with this application the students could compare solutions and visualize how the tractions and displacements are distributed on the typical problems solved in Strength of Materials and Elasticity courses. BEMAPEC provides a helpful tool for self-learning and a better understanding of the theoretical concepts.

The continuous evaluation allows for the assessment of the progressive assimilation of concepts and the competences that must be achieved in a course. There are several ways to implement such continuous evaluation system. We propose auto-evaluation tests as a valuable tool for the student to judge his level of knowledge. Furthermore, these tests are also used as a small part of the continuous evaluation process, encouraging students to learn the concepts seen in the course, as they have the feeling that the time dedicated to this study will have an assured reward, binge able to answer correctly the questions in the continuous evaluation exams. New technologies are a great aid to improve the auto-evaluation experience both for the students and the teachers. In this research work we have compared the results obtained in courses where auto-evaluation tests were provided against courses where they were not provided, showing how the tests improve a set of quality metrics in the results of the course.

Children with autism are characterized by impairments in communication, social interaction and information processing. This work presents a module to encourage children with autism to improve their social and communication skills, through a specially designed game-based approach. The humanoid robot NAO is utilized to autonomously engage with a child. The proposed module suggests a multiple role for the robot which can act as a teacher, as a toy and as a peer, through a successive set of joint activities. Overall observation encourages the utilization of NAO in the rehabilitation of children with autism.

We propose storytelling as a central tool in social robotics and its use in educational environments, either for the conventional classroom or for children with special needs. Storytelling is not only a way to convey a message to the audience, but it is also an excellent guide for interaction. Stories provide the context and can be used also to model the child attention and current state of knowledge of the topic, i.e. to achieve user modelling.

The Autism Spectrum Disorder (ASD) is a condition that can be highly challenging when facing interaction with others. One of the pivotal social skills that is not fully developed in ASD individuals is the joint-attention, due to high engagement in their own thoughts and emotions. Joint-attention is a base for important social behaviours. At the same time, promising results have arisen from studies on the interaction of social robots with ASD children, arguably because its low emotional evocation. In this paper we discuss an intervention proposal for a robot-interaction aiming to develop join-attention in ASD children. This intervention is based in two activities, each one can be graduated at three different level each, with an increasing level of socialization, in order to help develop this skills in a natural way.

This paper describes an educational experiment carried out by an Educational Innovation Project (EIP), developed in the field of Industrial Informatics during the biennium 2011/2013 at the Faculty of Engineering of Vitoria-Gasteiz (University of the Basque Country, UPV/EHU, Spain). In this paper the general situation regarding the European Higher Education Area (EHEA) at the start of that biennium, as well as the situation and specific problems that occurred in the field of Industrial Informatics at the Faculty are described. It was proposed to rectify the situation by the explicit formulation of ambitious objectives and the use of active learning methods, specifically by intragroup (between members of the same group) and intergroup (between members of different groups) cooperative learning. The paper includes the designing details of the proposed innovation carried out, the designed assessment, and the steps taken for the implementation in each one of the two years of the project. The results have been successful in the academic field since the specific and generic competences have been achieved and even from the point of view of the evaluation of teachers by students.

Since 2017–2018 academic year, the Bachelor’s Degree in Auto-motive Engineering has been implemented at the Faculty of Engineering of Vitoria-Gasteiz (Basque Country University, UPV/EHU) in dual format, based on the complementarity of the University and company learning environments. With this type of education the aim is to achieve greater motivation of the students and facilitate the work insertion by having greater contact with companies. In this way, companies are fully integrated with the agents who are responsible for the education of university students. The teaching institutions need to accredit Bachelor’s degrees or itineraries that are taught using this training format (dual training). In order to get this purpose, the university institutions adapt their studies in dual format to the quality requirements proposed by the state quality agencies of the university systems, based on objective evaluation criteria.

In this paper authors introduce an ongoing Educational Innovation Project (EIP) that is being carried out at the Faculty of Engineering of Vitoria-Gasteiz (Basque Country University, UPV/EHU, Spain). Since it is interesting from the point of view of universities, research institutes and even traditional companies, the aim of the project is to introduce the scientific culture in the Bachelor’s Degrees of the Faculty. The project has been granted in the call for Educational Innovation Projects 2018–2019 of the Educational Advisory Service, Vicerectorship for Teaching Quality and Innovation. In the paper, the key points of the granted project are explained in order to encourage similar actions, but since it has just started neither practical issues nor experiences can be still reported.