Search

This article proposes a new method for estimating the extrinsic calibration parameters between any pair of multibeam LiDAR sensors on a vehicle. Unlike many state-of-the-art works, this method does not use any calibration pattern or reflective marks placed in the environment to perform the calibration; in addition, the sensors do not need to have overlapping fields of view. An iterative closest point (ICP)-based process is used to determine the values of the calibration parameters, resulting in better convergence and improved accuracy. Furthermore, a setup based on the car learning to act (CARLA) simulator is introduced to evaluate the approach, enabling quantitative assessment with ground-truth data. The results show an accuracy comparable with other approaches that require more complex procedures and have a more restricted range of applicable setups. This work also provides qualitative results on a real setup, where the alignment between the different point clouds can be visually checked. The open-source code is available at https://github.com/midemig/pcd_calib . This work was supported in part by the Madrid Government (Comunidad de Madrid-Spain) under the Multiannual Agreement with UC3M ("Fostering Young Doctors Research," APBI-CM-UC3M) in the context of the V PRICIT (Research and Technological Innovation Regional Program); and in part by the Spanish Government through Grants ID2021-128327OA-I00 and TED2021-129374A-I00 funded by MCIN/AEI/10.13039/501100011033 and by the European Union NextGenerationEU/PRTR.

The Google's Loon[TM] initiative aims at covering rural or underdeveloped areas via fleets of high-altitude balloons supporting LTE connectivity. But how effective and stable can be the coverage provided by a network deployed via propulsion-free balloons, floating in the sky, and only loosely controllable through altitude variations? To provide some insights on the relevant performance and trade-offs, in this paper we gather real-world data from publicly available flight tracking services, and we analyze coverage and service stability in three past deployment scenarios. Besides employing a variety of metrics related to spatial and temporal coverage, we also assess service continuity, by also leveraging recently proposed “meaningful availability” metrics. While our analyses show that balloons are certainly a cost-effective way to provide a better-than-nothing and delay-tolerant service, there is yet no empirical evidence that an increase in the number of overlapping balloons may be rewarded with a substantial performance increase - in other words, we suspect that guaranteeing coverage and service stability levels comparable to that of a terrestrial cellular network is a challenging goal. Publicado

Progress in the design of functional perovskite oxides relies on advances in density functional theory (DFT) methods to efficiently and effectively model complex systems composed of several transition-metal ions. This work reports the application of DFT methods to investigate the electronic structure of the YSr2Cu2FeO8−δ (0 < δ < 1) family in which the insulating, metal, or superconducting behaviors and even anion conductivity can be tuned by modifying the oxygen content. In particular, we assess the performance of the generalized gradient approximation (GGA), its Hubbard-U correction (GGA + U), and the strongly constrained and appropriately normed (SCAN) to model the metallic (idealized YSr2Cu2FeO8) and insulating (idealized YSr2Cu2FeO7) phases of the system. The analysis of the DFT results is supported by DC resistivity measurements that denote the metal character of the synthesized YSr2Cu2FeO7.86 and the semiconducting character of YSr2Cu2FeO7.08 prepared under reducing conditions. In addition, the band gap of YSr2Cu2FeO7.08, in the range of 0.73−1.2 eV, has been extracted from diffuse reflectance spectroscopy (DRS). While the three methodologies (GGA, GGA + U, SCAN) permit the reproduction of the crystal structures of the synthetized oxides (determined here in the case of YSr2Cu2FeO7.08 by neutron powder diffraction (NPD)), the SCAN emerges as the only one capable to predict the basic electronic and magnetic properties across the YSr2Cu2FeO8−δ (0 < δ < 1) series. The picture that emerges for the metal (δ = 0) to insulating (δ = 1) transition is the one in which oxygen vacancies contribute electrons to the filling of the Cu/Fe-3dx2−y2 states of the conduction band. These results validate the SCAN functional for future DFT investigations of complex functional oxides that combine severaltransition metals.

Hybrid platforms combining multicore Central Processing Units (CPU) with many-core hardware accelerators such as Graphic Processing Units (GPU) can be smartly exploited to provide efficient parallel implementations of wireless communication algorithms for Fifth Generation (5G) and beyond systems. Massive Multiple-Input Multiple-Output (MIMO) systems are a key element of the 5G standard, involving several tens or hundreds of antenna elements for communication. Such a high number of antennas has a direct impact on the computational complexity of some MIMO signal processing algorithms. In this work, we focus on the channel estimation stage. In particular, we develop a parallel implementation of a recently proposed MIMO channel estimation algorithm. Its performance in terms of execution time is evaluated both in a multicore CPU and in a GPU. The results show that some computation blocks of the algorithm are more suitable for multicore implementation, whereas other parts are more efficiently implemented in the GPU, indicating that a hybrid CPU-GPU implementation would achieve the best performance in practical applications based on the tested platform.

Quantum computing poses fascinating challenges for current cryptography, threatening the security of many schemes and protocols widely used today. To adapt to this reality, the U.S. National Institute for Standards and Technology (NIST) is currently running a standardization process in search of post-quantum (classical, yet resistant to quantum attacks) cryptographic tools, focusing on signature schemes and key encapsulation mechanisms. Many of the competing proposals also include designs for two-party key exchange, which can be combined in different ways to fit scenarios involving (Formula presented.) parties, that is, yielding group key exchange protocols. However, very few implementations of such group protocols are available to practitioners, which face a non-trivial challenge when deciding how to implement a protocol for establishing secure group sessions in this new post-quantum scenario. With this in mind, the authors report on the implementation of a secure post-quantum group key exchange protocol in the so-called Quantum Random Oracle Model. The protocol decided to implement is based on a KEM called Kyber, which is one of the finalists of the NIST competition. Not only this group construction is the only one available in the literature using a NIST finalist, but also, among all post-quantum designs the authors are aware of, it uses this strongest security model (as, e.g. in other proposals, the adversarial interaction with the hash functions of the system is assumed to be exclusively classical). Furthermore, experimental evidence is provided supporting this choice in terms of performance, even if the number of involved entities is large (up to 2000). All data and code are publicly available. Ministerio de Economía y Competitividad. Grant Number: PID2019-109379RB-100 NATO Science for Peace and Security Programme. Grant Number: G5448

This is an Accepted Manuscript of an article published by Taylor & Francis Group in Soft matter on 29/03/2023, available online at: http://www.tandfonline.com/10.1039/d2sm01326f Cell migration is a physical process central to life. Among others, it regulates embryogenesis, tissue regeneration, and tumor growth. Therefore, understanding and controlling cell migration represent fundamental challenges in science. Specifically, the ability of cells to follow stiffness gradients, known as durotaxis, is ubiquitous across most cell types. Even so, certain cells follow positive stiffness gradients while others move along negative gradients. How the physical mechanisms involved in cell migration work to enable a wide range of durotactic responses is still poorly understood. Here, we provide a mechanistic rationale for durotaxis by integrating stochastic clutch models for cell adhesion with an active gel theory of cell migration. We show that positive and negative durotaxis found across cell types are explained by asymmetries in the cell adhesion dynamics. We rationalize durotaxis by asymmetric mechanotransduction in the cell adhesion behavior that further polarizes the intracellular retrograde flow and the protruding velocity at the cell membrane. Our theoretical framework confirms previous experimental observations and explains positive and negative durotaxis. Moreover, we show how durotaxis can be engineered to manipulate cell migration, which has important implications in biology, medicine, and bioengineering. P.S acknowledges support from the Spanish Ministry of Economy and Competitiveness (Grant No: PID2019-110949GB-I00) and the Generalitat de Catalunya (Grant No: 2017-SGR-1278). C.V. was supported by the Generalitat de Catalunya (FI grant). P.S and C.V. acknowledge support from the European Commission (Grant No. H2020-FETPROACT-01-2016-731957). We also thank W. Mirza, J.J. Muñoz, and M. Arroyo for discussion.

The main research motivation of this article is the fight against gender-based violence and achieving gender equality from a technological perspective. The solution proposed in this work goes beyond currently existing panic buttons, needing to be manually operated by the victims under difficult circumstances. Instead, Bindi, our end-to-end autonomous multimodal system, relies on artificial intelligence methods to automatically identify violent situations, based on detecting fear-related emotions, and trigger a protection protocol, if necessary. To this end, Bindi integrates modern state-of-the-art technologies, such as the Internet of Bodies, affective computing, and cyber-physical systems, leveraging: 1) affective Internet of Things (IoT) with auditory and physiological commercial off-the-shelf smart sensors embedded in wearable devices; 2) hierarchical multisensorial information fusion; and 3) the edge-fog-cloud IoT architecture. This solution is evaluated using our own data set named WEMAC, a very recently collected and freely available collection of data comprising the auditory and physiological responses of 47 women to several emotions elicited by using a virtual reality environment. On this basis, this work provides an analysis of multimodal late fusion strategies to combine the physiological and speech data processing pipelines to identify the best intelligence engine strategy for Bindi. In particular, the best data fusion strategy reports an overall fear classification accuracy of 63.61% for a subject-independent approach. Both a power consumption study and an audio data processing pipeline to detect violent acoustic events complement this analysis. This research is intended as an initial multimodal baseline that facilitates further work with real-life elicited fear in women. This work was supported in part by the Department of Research and Innovation of Madrid Regional Authority, in the EMPATIA-CM Research Project (Reference Y2018/TCS-5046) funded by MCIN/AEI/10.13039/501100011033 under Grant PDC2021-121071-I00; in part by the European Union NextGenerationEU/PRTR in part by the Spanish Ministry of Universities with the FPU under Grant FPU19/00448; and in part by the Madrid Government (Comunidad de Madrid-Spain) through the Multiannual Agreement with UC3M in the line of Excellence of University Professors (EPUC3M26), and in the context of the V PRICIT (Regional Programme of Research and Technological Innovation).

The collaboration of the real world and the virtual world, known as Digital Twin, has become a trend with numerous successful use cases. However, there are challenges mentioned in the literature that must be addressed. One of the most important issues is the difficulty of collaboration of Digital Twins due to the lack of standardization in their implementation. This article continues a previous work that proposed a generic architecture based on the FIWARE components to build Digital Twins in any field. Our work proposes the use of Linked Open Data as a mechanism to facilitate the communication of Digital Twins. We validate our proposal with a use case of an urban Digital Twin that collaborates with a parking Digital Twin. We conclude that Linked Open Data in combination with the FIWARE ecosystem is a real reference option to deploy Digital Twins and to enable the collaboration between Digital Twins.

This paper proposes the minimization of α-divergences for approximate inference in the context of deep Gaussian processes (DGPs). The proposed method can be considered as a generalization of variational inference (VI) and expectation propagation (EP), two previously used methods for approximate inference in DGPs. Both VI and EP are based on the minimization of the Kullback-Leibler divergence. The proposed method is based on a scalable version of power expectation propagation, a method that introduces an extra parameter α that specifies the targeted α-divergence to be optimized. In particular, such a method can recover the VI solution when α → 0 and the EP solution when α → 1. An exhaustive experimental evaluation shows that the minimization of α-divergences via the proposed method is feasible in DGPs and that choosing intermediate values of the α parameter between 0 and 1 can give better results in some problems. This means that one can improve the results of VI and EP when training DGPs. Importantly, the proposed method allows for stochastic optimization techniques, making it able to address datasets with several millions of instances, The authors gratefully acknowledge the use of the facilities of Centro de Computación Científica (CCC) at Universidad Autónoma de Madrid. The authors also acknowledge financial support from Spanish Plan Nacional I+D+i, Ministerio de Ciencia e Innovación, grant PID2019-106827GB-I00 / AEI / 10.13039/501100011033

The degradation of critical components inside large industrial assets, such as ball-bearings, has a negative impact on production facilities, reducing the availability of assets due to an unexpectedly high failure rate. Machine learning- based monitoring systems can estimate the remaining useful life (RUL) of ball-bearings, reducing the downtime by early failure detection. However, traditional approaches for predictive systems require run-to-failure (RTF) data as training data, which in real scenarios can be scarce and expensive to obtain as the expected useful life could be measured in years. Therefore, to overcome the need of RTF, we propose a new methodology based on online novelty detection and asymmetrical hidden Markov models (As-HMM) to work out the health assessment. This new methodology does not require previous RTF data and can adapt to natural degradation of mechanical components over time in data-stream and online environments. As the system is designed to work online within the electrical cabinet of machines it has to be deployed using embedded electronics. Therefore, a performance analysis of As-HMM is presented to detect the strengths and critical points of the algorithm. To validate our approach, we use real life ball-bearing data-sets and compare our methodology with other methodologies where no RTF data is needed and check the advantages in RUL prediction and health monitoring. As a result, we showcase a complete end-to-end solution from the sensor to actionable insights regarding RUL estimation towards maintenance application in real industrial environments. This study was supported partially by the Spanish Ministry of Economy and Competitiveness through the PID2019-109247GB-I00 project and by the Spanish Ministry of Science and Innovation through the RTC2019-006871-7 (DSTREAMS project). Also, by the H2020 IoTwins project (Distributed Digital Twins for industrial SMEs: a big-data platform) funded by the EU under the call ICT-11-2018- 2019, Grant Agreement No. 857191.

Bayesian networks are probabilistic graphical models that are commonly used to represent the uncertainty in data. The PyBNesian package provides an implementation for many different types of Bayesian network models and some variants, such as conditional Bayesian networks and dynamic Bayesian networks. In addition, the package can be easily extended with new components that can interoperate with those already implemented. Furthermore, the package also implements other related models such as kernel density estimation using OpenCL 1.2+ to enable GPU acceleration. PyBNesian is totally free and open-source under the MIT license.

Wind and solar energy forecasting have become crucial for the inclusion of renewable energy in electrical power systems. Although most works have focused on point prediction, it is currently becoming important to also estimate the forecast uncertainty. With regard to forecasting methods, deep neural networks have shown good performance in many fields. However, the use of these networks for comparative studies of probabilistic forecasts of renewable energies, especially for regional forecasts, has not yet received much attention. The aim of this article is to study the performance of deep networks for estimating multiple conditional quantiles on regional renewable electricity production and compare them with widely used quantile regression methods such as the linear, support vector quantile regression, gradient boosting quantile regression, natural gradient boosting and quantile regression forest methods. A grid of numerical weather prediction variables covers the region of interest. These variables act as the predictors of the regional model. In addition to quantiles, prediction intervals are also constructed, and the models are evaluated using different metrics. These prediction intervals are further improved through an adapted conformalized quantile regression methodology. Overall, the results show that deep networks are the best performing method for both solar and wind energy regions, producing narrow prediction intervals with good coverage.

Many real-world optimization problems involve two different subsets of variables: decision variables, and those variables which are not present in the cost function but constrain the solutions, and thus, must be considered during optimization. Thus, dependencies between and within both subsets of variables must be considered. In this paper, an estimation of distribution algorithm (EDA) is implemented to solve this type of complex optimization problems. A Gaussian Bayesian network is used to build an abstraction model of the search space in each iteration to identify patterns among the variables. As the algorithm is initialized from data, we introduce a new hyper-parameter to control the influence of the initial data in the decisions made during the EDA execution. The results show that our algorithm improves the cost function more than the expert knowledge does.

La falta de mujeres en las disciplinas científico-técnicas tiene no sólo múltiples causas, sino diferentes manifestaciones. En este trabajo, exploramos las temáticas en las que se centran los trabajos de fin de grado de Ingeniería Informática de la Universidade da Coruña como posible indicador de las diferencias en los roles jugados por mujeres y hombres en el mundo de la tecnología. En concreto, la cuantificación de trabajos fin de grado que se asocian con varios ámbitos asociados a los cuidados determina que estas temáticas son poco frecuentes, y el análisis de la autoría y tutorización revela que se registran diferencias entre mujeres y hombres en ambos roles, si bien de diferente magnitud.

This paper presents a new class of Bayesian networks called hybrid semiparametric Bayesian networks, which can model hybrid data (discrete and continuous data) by mixing parametric and nonparametric estimation models. The parametric estimation models can represent a conditional linear Gaussian relationship between variables, while the nonparametric estimation model can represent other types of relationships, such as non-Gaussian and nonlinear relationships. This new class of Bayesian networks generalizes the conditional linear Gaussian Bayesian networks, including them as a special case. In addition, we describe a learning procedure for the structure and the parameters of our proposed type of Bayesian network. This learning procedure finds the best combination of parametric and nonparametric models automatically from data. This requires the definition of a cross-validated score. We also detail how new data can be sampled from a hybrid semiparametric Bayesian network, which in turn can be useful to solve other related tasks, such as inference. Furthermore, we intuitively relate our proposal with adaptive kernel density estimation models. The experimental results show that hybrid semiparametric Bayesian networks are a valuable contribution when dealing with data that do not meet the parametric assumptions that are expected for other models, such as conditional linear Gaussian Bayesian networks. We include experiments with synthetic data and real-world data from the UCI repository which demonstrate the good performance and the ability to extract useful information about the relationship between the variables in the model.

In this paper, we focus on developing automatic assessment (AA) for a topic that has some difficulties in its practical assessment: object oriented programming (OOP). For evaluating that the OOP principles have been correctly applied to a real application, we use unit testing. In this paper, we focus on prioritizing that the students understand and apply correctly complex OOP principles and that they design properly the classes (including their relationships). In addition, we focus on the Python programming language rather than the typical previous works' focus in this area. Thus, we present a real case study of a practical assignment, in which the students have to implement characters for a video game. This assignment has the particularities and advantages that it is incremental and that it applies all four OOP principles within a single assignment. We also present its solution with the UML class diagram description. Furthermore, we provide unit testing for this case study and give general advice for generalizing the unit tests to other real case scenarios. Finally, we corroborate the effectiveness of our approach with positive student evaluations

The Internet of things is a paradigm with the goal of creating customizable goods and services based on user experience. This paradigm has been applied in industrial environments generating what is called the industrial Internet of things. This new paradigm measures industrial assets continuously. The collected information is processed to extract data insights regarding the assets status or health. Depending on the asset status, maintenance policies can be planned to prevent the assets failure or degradation. To elaborate such insights, artificial intelligence models are usually applied to learn and predict industrial data patterns and behaviors. Nonetheless, in many cases, computational and reliability restrictions are present, and fast and explainable models are required to satisfy the industry needs. Hidden Markov models (HMMs) are statistical models that are capable of learning data patterns and detect non-stationary behavior in data. When HMMs are compared to other models, HMMs are economic, explainable and reliable. They are economic because their learning and inference algorithms can run in a reasonable time without the need for graphic cards or other power-intensive computing devices. They are explainable since all the learned parameters are interpretable from a probabilistic and area-of-knowledge points of view. They are reliable, because, if a mistake is committed by the model, it is possible to detect and infer its causes from the models parameters and structure. Due to the previous discussion, the motivation and results of this thesis aim to extend theoretically current HMMs, to make them more relevant, general and useful for industrial applications. For all the proposed models, the expectation-maximization algorithm was used for the learning phase. The first contribution appears in Chapter 4, where context-specific Bayesian networks were used to model the emission probabilities of continuous variables. That model is referred as AsLG-HMM since linear Gaussian Bayesian networks were used. The model was compared to a mixture of Gaussian HMM, where improvements in log-likelihoods by the proposed model were observed in both synthetic data and real data from ball-bearings. Nonetheless, such model was further developed in Chapter 5, where autoregressive values of the observable variables were considered in the context-specific Bayesian networks. This model is referred to AR-AsLG-HMM. In this case, the model was studied with further mathematical rigor. Also, a forward greedy algorithm was proposed to discover structures of Bayesian networks for the emission probabilities. The model was tested with synthetic and real data incoming from air quality and ball-bearing data. For this model, several types of HMMs were used for comparison. The learning times were also considered for evalua tion. The proposed model showed improvements in log-likelihood with fair learning times, and additional data insights were provided due to the learned Bayesian networks. As comment, AR-AsLG-HMM served as a cornerstone model for other contributions in this thesis. In Chapter 6, AR-AsLG-HMM was endowed with feature saliencies to enable the model to perform an embedded feature selection procedure. This implies that the model during its learning procedure determined the relevant features. This model is referred to FS-AsHMM. In this case, the model was compared to other HMMs with feature saliencies. Synthetic data and real data from ball-bearings and cameras with face- expressions data were used for validation. The model obtained better results regarding expression recognition and detection of non-relevant features. The previous contributions were focused to offline analysis. Nevertheless, this thesis is focused on working in industrial environments, where data-streams are generated and the models are expected to adapt to changes in data. To address such issue, in Chapter 7 the AR-AsLG-HMM was adapted to be used in data-stream and perform continuous learning. Novel-concept detection techniques were used to determine when new unobserved patterns appeared. Based on the data-insights of the AR-AsLG-HMM from the data-stream, a healthindex and a regression model were proposed to determine the health status and remaining useful life of ball-bearings. Two datasets were used to validate the proposed methodology: open access datasets with ball-bearings which are run to failure, and a ball-bearing testbed from a company promoting the thesis, Aingura IIoT. Additionally, in collaboration with the Barcelona Supercomputer Center, the methodology code was optimized to be embedded into edge devices and use it in real life applications. The methodology was compared to others in the state of the art. It obtained better results in terms of health estimation, and fair results regarding the remaining useful life prediction. Next, in Chapter 8 a feature saliency model for HMMs was adapted to determine relevant harmonics of ball-bearings data in online environments. However, this study was a preliminary work for what was done in Chapter 9, where local feature saliencies were applied on AR-AsLG-HMMs. This model is referred to LFS-AsHMM. This model was adapted to be used in data-streams with novel-concept detection techniques to keep track of the evolution of relevant features. This model updated the relevant features only when the data needed it. Synthetic and real open access data from ball-bearings was used for validation. The model was compared to other strategies and methodologies that perform feature selection in data-streams. However, these strategies did the feature selection whenever a new instance arrived and not when needed. Unfortunately, this model did not get to be implemented into edge devices during the writing of this thesis. Finally, the proposed models assume linear Gaussian data and if such assumption fails, the models are no longer valid. To address such problem, in Chapter 10, the ideas used on AR-AsLG-HMM were imposed over HMMs with non-parametric emission probabilities, more precisely, kernel density estimations were used to approximate the emission probabilities, and the estimations depended on context-specific Bayesian networks. The proposed model is referred to KDE-AsHMM. The proposed model is validated using synthetic non-linear Gaussian data and open access real data from sound recognition problems and drill milling processes. The model showed improvements in likelihood and sound recognition accuracy when compared to other HMMs. Nonetheless, the learning times and computational resources were high demanding. At the end of the thesis, in Chapter 11, the corresponding conclusions, final remarks and future research lines were proposed. RESUMEN El internet de las cosas es un paradigma con el objetivo de crear bienes y servicios perzonalizados basados en la experiencia del usuario. Este paradigma se ha aplicado en entornos industriales generando lo que se denomina el Internet de las cosas industriales. Este nuevo paradigma mide los activos industriales de forma continua. La información recopilada se procesa para extraer información sobre el estado o la salud de los activos. Según el estado de los activos, se pueden planificar políticas de mantenimiento para evitar fallas o degradación de los activos. Para elaborar tales conocimientos, modelos de inteligencia artificial se aplican para aprender y predecir patrones y comportamientos de datos industriales. No obstante, en muchos casos existen restricciones computacionales y de confiabilidad, y se requieren modelos rápidos y explicables para satisfacer las necesidades de la industria. Los modelos ocultos de Markov (HMM) son modelos estadísticos que son capaces de aprender patrones de datos y detectar comportamientos no estacionarios en los datos. Cuando los HMMs se comparan con otros modelos, los HMMs son económicos, explicables y confiables. Son económicos porque sus algoritmos de aprendizaje e inferencia pueden ejecutarse en un tiempo razonable sin necesidad de tarjetas gráficas u otros dispositivos informáticos que consumen mucha energía. Son explicables ya que todos los parámetros aprendidos son interpretables desde un punto de vista probabilístico y de área de conocimiento. Son confiables, puesto que si el modelo comete un error, es posible detectar e inferir sus causas a partir de los parámetros y la estructura del modelo. Debido a la discusión anterior, la motivación y los resultados de esta tesis tienen como objetivo extender los HMM teóricamente actuales, para hacerlos más relevantes, generales y útiles para aplicaciones industriales. Para todos los modelos propuestos se utilizó el algoritmo de maximización de expectativas para la fase de aprendizaje. La primera contribución aparece en el Capítulo 4, donde se utilizaron redes bayesianas específicas del contexto para modelar las probabilidades de emisión de variables continuas. Ese modelo se conoce como AsLG-HMM ya que se utilizaron redes lineales gaussianas bayesianas. El modelo se comparó con una mezcla de Gaussian HMM, donde se observaron mejoras en las probabilidades logarítmicas del modelo propuesto tanto en datos sintéticos como en datos reales de rodamientos de bolas. No obstante, dicho modelo se desarrolló más en el Capítulo 5, donde se consideraron los valores autorregresivos de las variables observables en las redes bayesianas específicas del contexto. Este modelo se denomina AR-AsLG-HMM. En este caso, el modelo fue estudiado con mayor rigor matemático. Además, se propuso un algoritmo voraz directo para descubrir estructuras de redes bayesianas para las probabilidades de emisión. El modelo se probó con datos sintéticos y reales provenientes de la calidad del aire y datos de cojinetes de bolas. Para este modelo, se usaron varios tipos de HMM para comparar. Los tiempos de aprendizaje también fueron considerados para la evaluación. El modelo propuesto mostró mejoras en la probabilidad de registro con tiempos de aprendizaje justos, y se proporcionaron conocimientos de datos adicionales debido a las redes bayesianas aprendidas. Como comentario, AR-AsLG-HMM sirvió como modelo fundamental para otras contribuciones en esta tesis. En el Capítulo 6, se dotó a AR-AsLG-HMM con variables destacadas para permitir que el modelo realice un procedimiento de selección de variables incorporado. Esto implica que el modelo durante su procedimiento de aprendizaje determinó las variables relevantes. Este modelo se denomina FS-AsHMM. En este caso, el modelo se comparó con otros HMM con variables sobresalientes. Para la validación se utilizaron datos sintéticos y datos reales de rodamientos de bolas y cámaras con datos de expresiones faciales. El modelo obtuvo mejores resultados en cuanto al reconocimiento de expresiones y detección de variables no relevantes. Las contribuciones anteriores estaban enfocadas al análisis fuera de línea. Sin embargo, esta tesis se centra en trabajar en entornos industriales, donde se generan flujos de datos y se espera que los modelos se adapten a los cambios en los datos. Para abordar este problema, en el Capítulo 7, el AR-AsLG-HMM se adaptó para usarse en flujo de datos y realizar un aprendizaje continuo. Se utilizaron técnicas de detección de conceptos novedosos para determinar cuándo aparecían nuevos patrones no observados. Con base en los conocimientos de datos del AR-AsLG-HMM del flujo de datos, se propusieron un índice de salud y un modelo de regresión para determinar el estado de salud y la vida útil restante de los rodamientos de bolas. Se utilizaron dos conjuntos de datos para validar la metodología propuesta: conjuntos de datos de acceso abierto con rodamientos de bolas que funcionan hasta el fallo y un banco de pruebas de rodamientos de bolas de la empresa promotora de la tesis, Aingura IIoT. Además, en colaboración con el Barcelona Supercomputing center, se optimizó el código de la metodología para integrarlo en edge devices y usarlo en aplicaciones de la vida real. La metodología fue comparada con otras en el estado del arte. Obtuvo mejores resultados en cuanto a la estimación de la salud, y resultados regulares en cuanto a la predicción de la vida útil remanente. A continuación, en el Capítulo 8, se adaptó un modelo de prominencia de variable para HMM para determinar los armónicos relevantes de los datos de rodamientos en entornos en línea. Sin embargo, este estudio fue un trabajo preliminar para lo que se hizo en el Capítulo 9, donde se aplicaron las prominencias de variables locales en AR-AsLG-HMM. Este modelo se denomina LFS-AsHMM. Este modelo se adaptó para usarse en flujos de datos con técnicas de detección de conceptos novedosos para realizar un seguimiento de la evolución de las variables relevantes. Este modelo actualizó las variables relevantes solo cuando los datos lo necesitaban. Para la validación se utilizaron datos de acceso abierto sintéticos y reales de rodamientos de bolas. El modelo se comparó con otras estrategias y metodologías que realizan la selección de variables en flujos de datos. Sin embargo, estas estrategias hacían la selección de funciones cada vez que llegaba una nueva instancia y no cuando era necesario. Desafortunadamente, este modelo no llegó a implementarse en dispositivos de Edge durante la redacción de esta tesis. Finalmente, los modelos propuestos asumen datos gaussianos lineales y si tal suposición falla, los modelos ya no son válidos. Para abordar tal problema, en el Capítulo 10, las ideas utilizadas en AR-AsLG-HMM se impusieron sobre los HMM con probabilidades de emisión no paramétricas, más precisamente, se utilizaron estimaciones de densidad kernel para aproximar las probabilidades de emisión, y las estimaciones dependían de redes bayesianas específicas del contexto. El modelo propuesto se refiere a KDE-AsHMM. El modelo propuesto se valida utilizando datos gaussianos no lineales sintéticos y datos reales de acceso abierto de problemas de reconocimiento de sonido y procesos de fresado de perforación. El modelo mostró mejoras en la probabilidad y la precisión del reconocimiento de sonido en comparación con otros HMM. No obstante, los tiempos de aprendizaje y los recursos computacionales fueron muy exigentes. Al final de la tesis, en el Capítulo 11, se propusieron las correspondientes conclusiones, comentarios finales y futuras líneas de investigación.

Modelling the biology behind animal behaviour has attracted great interest in recent years. Nevertheless, neuroscience and artificial intelligence face the challenge of representing and emulating animal behaviour in robots. Consequently, this paper presents a biologically inspired motivational model to control the biological functions of autonomous robots that interact with and emulate human behaviour. The model is intended to produce fully autonomous, natural, and behaviour that can adapt to both familiar and unexpected situations in human–robot interactions. The primary contribution of this paper is to present novel methods for modelling the robot’s internal state to generate deliberative and reactive behaviour, how it perceives and evaluates the stimuli from the environment, and the role of emotional responses. Our architecture emulates essential animal biological functions such as neuroendocrine responses, circadian and ultradian rhythms, motivation, and affection, to generate biologically inspired behaviour in social robots. Neuroendocrinal substances control biological functions such as sleep, wakefulness, and emotion. Deficits in these processes regulate the robot’s motivational and affective states, significantly influencing the robot’s decision-making and, therefore, its behaviour. We evaluated the model by observing the long-term behaviour of the social robot Mini while interacting with people. The experiment assessed how the robot’s behaviour varied and evolved depending on its internal variables and external situations, adapting to different conditions. The outcomes show that an autonomous robot with appropriate decision-making can cope with its internal deficits and unexpected situations, controlling its sleep–wake cycle, social behaviour, affective states, and stress, when acting in human–robot interactions. The research leading to these results has received funding from the projects: Robots Sociales para Estimulación Física, Cognitiva y Afectiva de Mayores (ROSES), RTI2018-096338-B-I00, funded by the Ministerio de Ciencia, Innovación y Universidades; Robots sociales para mitigar la soledad y el aislamiento en mayores (SOROLI), PID2021-123941OA-I00, funded by Agencia Estatal de Investigación (AEI), Spanish Ministerio de Ciencia e Innovación. This publication is part of the R&D&I project PLEC2021-007819 funded by MCIN/AEI/10.13039/501100011033 and by the European Union NextGenerationEU/PRTR.

The most challenging application of post-quantum cryptography (PQC) is the distribution of provably secure cryptographic keys and quantum-safe algorithms. Public-key Encryption (PKE), Key-establishment Mechanisms (KEM), and Digital Signature Algorithms (DSA) are necessary cryptographic mechanisms to use common encryption communication through insecure or untrusted networks. This paper proposes a provably quantum-resilient secure key agreement mechanism based on the most promising lattice-based cryptographic scheme and the Module Learning-with-Rounding (MLWR) problem. The main contribution of this paper is to construct a modified framework on classical computers for study and test-bed purposes.

Cloud computing has emerged as an attractive and fastest-growing technology. The Cloud model encompasses a highly scalable infrastructure environment and guarantees the daily business processes for data-driven business cases protected by quantum vulnerable cryptography and no preserving of data privacy. This paper proposes a practical quantum-safe approach to guarantee data privacy even if a quantum computer can break classical public-key exchange algorithms. We present a quantum-safe cloud use based on post-quantum cryptographic and privacy-focused applications via homomorphic encryption and secure multiparty computation.

Precision agriculture in the realm of the Internet of Things is characterized by the collection of data from multiple sensors deployed on the farm. These data present a spatial, temporal, and semantic characterization, which further complicates the performance in the management and implementation of models and repositories. In turn, the lack of standards is reflected in insufficient interoperability between management solutions and other non-native services in the framework. In this paper, an innovative system for spatio-temporal semantic data management is proposed. It includes a data query system that allows farmers and users to solve queries daily, as well as feed decision-making, monitoring, and task automation solutions. In the proposal, a solution is provided to ensure service interoperability and is validated against two European smart farming platforms, namely AFarCloud and DEMETER. For the evaluation and validation of the proposed framework, a neural network is implemented, fed through STSDaMaS for training and validation, to provide accurate forecasts for the harvest and baling of forage legume crops for livestock feeding. As a result of the evaluation for the training and execution of neural networks, high performance on complex spatio-temporal semantic queries is exposed. The paper concludes with a distributed framework for managing complex spatio-temporal semantic data by offering service interoperability through data integration to external agricultural data models. Graphical Abstract

186 p. [ES] Los dispositivos hápticos permiten percibir mediante el sentido del tacto propiedades de objetos virtuales. En base al tipo de realimentación que ofrecen se pueden diferenciar entre dispositivos táctiles o dispositivos de realimentación de fuerzas. Los de realimentación de fuerzas son los considerados más realistas ya que responden a la interacción que el usuario realiza con el sistema y transmiten propiedades de los objetos como su forma geométrica, dureza y peso. Para ofrecer esta realimentación deben contar con enlaces mecánicos capaces de actuar sobre el efector final que manipula el usuario. Estas uniones mecánicas suponen la principal limitación de estos dispositivos, ya que el espacio de trabajo disponible se encuentra siempre limitado por las dimensiones físicas de los brazos. Esta limitación hace que sean necesarias técnicas de interacción que permitan utilizar estos dispositivos para interaccionar hápticamente con escenas de dimensiones mayores a las de su espacio de trabajo. La revisión de la literatura llevada a cabo identifica varias técnicas específicamente diseñadas para abordar esta limitación. Sin embargo, la revisión también evidencia que ninguna de ellas permite una integración sencilla de estos dispositivos en experiencias de realidad virtual y que los usos que se han hecho de los dispositivos hápticos en estos entornos inmersivos han contado con una interacción limitada al no aplicar técnicas específicas. Para dar solución al problema identificado se ha desarrollado un primer modelo de interacción denominado zoom háptico que aborda el problema del espacio de trabajo limitado de los dispositivos hápticos de realimentación de fuerzas de escritorio. Este modelo permite, mediante un mapeado continuo del espacio de trabajo del dispositivo y del área que visualiza el usuario, tanto desplazarse por toda la escena con libertad de movimiento como centrarse en un área de interés realizando una serie de ampliaciones visuales y hápticas. Los resultados de la evaluación realizada comparando la nueva técnica con técnicas clásicasmuestran que la nueva técnica es tan fácil de utilizar como el mapeado directo y que ofrece tanta precisión como el clutching. Se ha desarrollado un segundo modelo de interacción centrado en permitir el uso de los dispositivos hápticos de realimentación de fuerzas de escritorio como principales mecanismos de interacción en experiencias de realidad virtual. Este nuevo modelo se denomina realidad háptica y su funcionamiento se basa en definir una representación virtual del espacio de trabajo del dispositivo llamada área de acción. La técnica permite la interacción háptica con los elementos que se encuentren dentro del área de acción y, para dotar al usuario de mayor precisión, integra la técnica zoom háptico que permite modificar el tamaño del área de acción. Para poder interaccionar con los objetos fuera del área de acción, el modelo contempla un modo recolocación que permite reposicionar dicha área en cualquier lugar de la escena. Esta técnica se ha implementado en el simulador médico ParaVR y este ha sido probado por un conjunto de paramédicos de servicio de emergencias Wales Ambulance Service NHS Trust, los cuales valoraron positivamente tanto el simulador como la técnica. La evaluación formal de la técnica realidad háptica se ha llevado a cabo en una comparativa con la técnica de la burbuja y los resultados han mostrado que la nueva técnica es superior a la de la burbuja en varios aspectos, colocándola como la técnica de referencia a la hora de integrar los dispositivos hápticos de realimentación de fuerzas de escritorio en experiencias inmersivas de realidad virtual. [EN] Haptic devices make it possible to perceive properties of virtual objects by means of the sense of touch. Based on the type of feedback they provide, a distinction can be made between haptic devices and force-feedback devices. Force feedback devices are considered the most realistic since they respond to the user’s interaction with the system and transmit object properties such as geometric shape, hardness and weight. To provide this feedback they must have mechanical links capable of acting on the end effector manipulated by the user. These mechanical links are the main limitation of these devices, since the available working space is always limited by the physical dimensions of the arms. This limitation makes necessary interaction techniques that allow the use of these devices to interact haptically with scenes of larger dimensions than those of their workspace. The literature review conducted identifies several techniques specifically designed to address this limitation. However, the review also evidences that none of them allow for easy integration of these devices into virtual reality experiences and that the uses that have been made of haptic devices in these immersive environments have featured limited interaction by not applying specific techniques. To provide a solution to the identified problem, a first interaction model called haptic zoom has been developed that addresses the problem of the limited workspace of forcefeedback desktop haptic devices. This model allows, through a continuous mapping of the device’s workspace and the area viewed by the user, both to move freely throughout the scene and to focus on an area of interest by performing a series of visual and haptic zooms. The results of the evaluation performed comparing the new technique with classical techniques shows that the new technique is as easy to use as direct mapping and offers as much accuracy as clutching. A second interaction model has been developed that focuses on enabling the use of force-feeback desktop haptic devices as the primary interaction mechanisms in virtual reality experiences. This new model is called haptic reality and its operation is based on defining a virtual representation of the device workspace called the action area. The technique allows haptic interaction with the elements inside the action area and, to provide the user with greater precision, it integrates the haptic zoom technique, which allows the size of the action area to be modified. In order to interact with objects outside the action area, the model includes a relocating mode that allows repositioning the action area anywhere in the scene. This technique has been implemented in the ParaVR medical simulator and has been tested by a set of emergency service paramedics from the Wales Ambulance Service NHS Trust, who rated both the simulator and the technique positively. Formal evaluation of the haptic reality technique has been carried out in a comparison with the bubble technique and the results have shown that the new technique is superior to the bubble technique in several respects, placing it as the benchmark technique when integrating force-feedback desktop haptic devices into immersive virtual reality experiences.

280 p. [ES] La navegación autónoma se presenta como la mayor revolución en la historia del transporte, debido a que los últimos avances científicos conseguidos avalan décadas de investigación en un sector con el potencial para solucionar los graves problemas del actual sistema de transporte. Sin embargo, la existencia de un elevado número de vehículos autónomos navegando en un mismo escenario plantea cuestiones y retos muy relevantes que aún no han sido resueltos. El trasfondo de esta cuestión es la automatización del tráfico en su conjunto. Las diferencias entre la automatización completa de un vehículo y la automatización del tráfico son muchas y notables. La investigación realizada en esta tesis analiza en profundidad la automatización del tráfico a nivel nacional, los retos científicos y tecnológicos, así como las implicaciones y consecuencias legislativas, económicas y sociales. Se presenta un modelo conceptual de sistema automatizado de transporte, diseñado para una movilidad en la que los sistemas autónomos reemplazan a las personas en la navegación. Es un sistema fundamentado en tres elementos principales que son: la gestión automatizada del tráfico, ejecutada por los Centros de Control del Tráfico (CCT), los ingenios autónomos, encargados de realizar las misiones propias del transporte, y una red de comunicaciones dedicada a la movilidad automatizada, o Red S3. La combinación de estos elementos permite la generación de un nuevo espacio radioeléctrico, el denominado Espacio Automatizado Multicontexto (EAMu), con importantes implicaciones operativas y comerciales. El sistema de transporte automatizado se caracteriza por la intermodalidad, integrando la navegación autónoma terrestre, aérea y marina. Automatizar el tráfico de un país, en este caso España, tiene importantes connotaciones jurídicas y administrativas. Se analizan los paradigmas fundamentales que definirán la nueva legislación que será necesaria para una movilidad automatizada. El proceso legislativo se estructurará en diferentes fases, en las que se irán escalando las acciones regulatorias. Se evalúa la estructura administrativa del trasporte actual en España, identificando los organismos y entidades que deberán ejecutar estas acciones legislativas. El impacto de la automatización en la movilidad también será muy importante. Se presenta un estudio en base a datos de movilidad reales de una ciudad de referencia, y los resultados obtenidos muestran la magnitud de esta revolución. La irrupción de los sistemas autónomos inteligentes permite redefinir las funcionalidades de los ingenios. En situaciones ideales se podría reducir el parque móvil por encima del 90%, satisfaciendo completamente la demanda de movilidad. Además, se puede llegar a duplicar la velocidad media de circulación en los núcleos urbanos. De igual modo, el impacto en la economía nacional es elevadísimo. Se han analizado los sectores más afectados, que representan el 38,16% del PIB de España. Las consecuencias de la automatización del tráfico en cada uno de ellos son muy grandes, en unos casos positivas y en otros negativas. La automatización del tráfico permite un modelo de movilidad sostenible, eficiente y seguro. Redefine los patrones y reglas de tráfico, pasando del tráfico por lotes al tráfico de flujos. Y también proporciona una movilidad bajo demanda e intermodal, dando lugar a una nueva estructura económica y permitiendo la recuperación de las ciudades para las personas. [EN] Autonomous navigation is presented as the greatest revolution in the history of transportation because the latest scientific advances achieved endorse decades of research in a sector with the potential to solve the serious problems of the current transportation system. However, the existence of a large number of autonomous vehicles navigating in the same scenario raises very relevant issues and challenges that have not yet been resolved. The background to this issue is the traffic automation. The differences between achieving fully autonomous vehicles and achieving traffic automation are many and significant. The research conducted in this thesis analyzes in depth the automation of the traffic at the national level, the scientific and technological challenges, as well as the legislative, economic and social implication and consequences. It presents a conceptual model of an automated transportation system designed for a mobility model in which autonomous systems replace people in navigation. It is a system based on three main elements: automated traffic management, executed by Traffic Control Centers (TCCs), autonomous vehicles, in charge of performing transport missions, and a communications network dedicated to autonomous mobility, or S3 Network. The combination of these elements allows the generation of a new radio space, the so-called Multidimensional Automated Space (MAS), with important operational and commercial implications. This automated transportation system is characterized by intermodality, integrating autonomous land, air and marine navigation. Automating a country’s traffic, in this case Spain, has important legal and administrative connotations. The fundamental paradigms that will define the new legislation that will be necessary for automated mobility are analyzed. The legislative process will be structured in different phases in which regulatory actions will be scaled. The current administrative structure of transport in Spain is evaluated, identifying the organizations and entities that will have to execute this legislative actions. The impact of automation on mobility will also be very important. A study is presented in a reference city based on real mobility data, and the results obtained show the magnitude of this revolution. The advent of intelligent autonomous systems makes it possible to redefine the functionalities of the vehicles. In ideal situations, the vehicle fleet could be reduced by more than 90% to meet the demand for mobility, doubling the average speed of traffic in urban areas. Similarly, the impact on the national economy is extremely high. The most affected sectors have been analyzed, representing 38,16% of Spain`s GDP. The consequences of traffic automation in each of them are very important, in some cases very positive an in others very negative. Traffic automation enables a sustainable, efficient and safe mobility model, redefining traffic patterns and rules, moving from batch traffic to flow traffic, providing on-demand and intermodal mobility, giving rise to a new economic structure and allowing the recovery of cities for people.

Esta tesis diseña y desarrolla un videojuego en realidad virtual para concienciar sobre la depresión a niñas y niños de entre 8 y 12 años. Se realiza una revisión bibliográfica de videojuegos de aplicación psicológica y de teorías de diseño para guiar este trabajo. Este análisis identifica los pilares, que son la narrativa, la jugabilidad y el aprendizaje, así como las necesidades de sus elementos para crear una experiencia divertida y educativa. El videojuego 3DC se produce desarrollando los modelos 3D y animaciones en Blender para después implementarlos y programar sus interacciones en Unreal. Por último, se realiza una validación del diseño y desarrollo con 21 participantes de 7 a 12 años, con los que se detecta un cambio en la percepción a partir de los 11 años. Los resultados obtenidos son positivos y permiten proponer mejoras para su futura implementación y validación del aprendizaje en un colegio., This thesis designs and develops a virtual reality serious game to raise awareness about depression among children between 8 and 12 years old. A bibliographic review of serious games with psychological applications and design theories is carried out to guide this work. This analysis identifies the pillars, which are narrative, playability and learning, as well as the needs of its elements to create a fun and educational experience. The 3DC video game is produced by developing the 3D models and animations in Blender and then implementing them and programming their interactions in Unreal. Finally, a validation of the design and development is performed with 21 participants from 7 to 12 years old, with whom a change in perception is detected from the age of 11. The results obtained are positive and allow to propose improvements for its future implementation and validation of learning in a school.

Both logic programming in general, and Prolog in particular, have a long and fascinating history, intermingled with that of many disciplines they inherited from or catalyzed. A large body of research has been gathered over the last 50 years, supported by many Prolog implementations. Many implementations are still actively developed, while new ones keep appearing. Often, the features added by different systems were motivated by the interdisciplinary needs of programmers and implementors, yielding systems that, while sharing the "classic" core language, and, in particular, the main aspects of the ISO-Prolog standard, also depart from each other in other aspects. This obviously poses challenges for code portability. The field has also inspired many related, but quite different languages that have created their own communities. This article aims at integrating and applying the main lessons learned in the process of evolution of Prolog. It is structured into three major parts. Firstly, we overview the evolution of Prolog systems and the community approximately up to the ISO standard, considering both the main historic developments and the motivations behind several Prolog implementations, as well as other logic programming languages influenced by Prolog. Then, we discuss the Prolog implementations that are most active after the appearance of the standard: their visions, goals, commonalities, and incompatibilities. Finally, we perform a SWOT analysis in order to better identify the potential of Prolog, and propose future directions along which Prolog might continue to add useful features, interfaces, libraries, and tools, while at the same time improving compatibility between implementations., 87 pages, 2 figures. This article has been accepted for publication in Theory and Practice of Logic Programming (TPLP)

The use of Digital Twins in the industry has become a growing trend in recent years, allowing to improve the lifecycle of any process by taking advantage of the relationship between the physical and the virtual world. Existing literature formulates several challenges for building Digital Twins, as well as some proposals for overcoming them. However, in the vast majority of the cases, the architectures and technologies presented are strongly bounded to the domain where the Digital Twins are applied. This article proposes the FIWARE Ecosystem, combining its catalog of components and its Smart Data Models, as a solution for the development of any Digital Twin. We also provide a use case to showcase how to use FIWARE for building Digital Twins through a complete example of a Parking Digital Twin. We conclude that the FIWARE Ecosystem constitutes a real reference option for developing DTs in any domain.

Outlier detection, i.e., the task of detecting points that are markedly different from the data sample, is an important challenge in machine learning. When a model is built, these special points can skew the model training and result in less accurate predictions. Due to this fact, it is important to identify and remove them before building any supervised model and this is often the first step when dealing with a machine learning problem. Nowadays, there exists a very large number of outlier detector algorithms that provide good results, but their main drawbacks are their unsupervised nature together with the hyperparameters that must be properly set for obtaining good performance. In this work, a new supervised outlier estimator is proposed. This is done by pipelining an outlier detector with a following a supervised model, in such a way that the targets of the later supervise how all the hyperparameters involved in the outlier detector are optimally selected. This pipeline-based approach makes it very easy to combine different outlier detectors with different classifiers and regressors. In the experiments done, nine relevant outlier detectors have been combined with three regressors over eight regression problems as well as with two classifiers over another eight binary and multi-class classification problems. The usefulness of the proposal as an objective and automatic way to optimally determine detector hyperparameters has been proven and the effectiveness of the nine outlier detectors has also been analyzed and compared, The authors acknowledge financial support from the European Regional Development Fund and the Spanish State Research Agency of the Ministry of Economy, Industry, and Competitiveness under the projects TIN2016-76406-P (AEI/FEDER, UE) and PID2019-106827GB-I00. They also thank the UAM–ADIC Chair for Data Science and Machine Learning and gratefully acknowledge the use of the facilities of Centro de Computación Científica (CCC) at UAM

Part of a collection: Personalization and Adaptation in Human-Robot Interactive Communication. Adapting to dynamic environments is essential for artificial agents, especially those aiming to communicate with people interactively. In this context, a social robot that adapts its behaviour to different users and proactively suggests their favourite activities may produce a more successful interaction. In this work, we describe how the autonomous decision-making system embedded in our social robot Mini can produce a personalised interactive communication experience by considering the preferences of the user the robot interacts with. We compared the performance of Top Label as Class and Ranking by Pairwise Comparison, two promising algorithms in the area, to find the one that best predicts the user preferences. Although both algorithms provide robust results in preference prediction, we decided to integrate Ranking by Pairwise Comparison since it provides better estimations. The method proposed in this contribution allows the autonomous decision-making system of the robot to work on different modes, balancing activity exploration with the selection of the favourite entertaining activities. The operation of the preference learning system is shown in three real case studies where the decision-making system works differently depending on the user the robot is facing. Then, we conducted a human-robot interaction experiment to investigate whether the robot users perceive the personalised selection of activities more appropriate than selecting the activities at random. The results show how the study participants found the personalised activity selection more appropriate, improving their likeability towards the robot and how intelligent they perceive the system. query Please check the edit made in the article title. The research leading to these results has received funding from the projects Robots Sociales para Estimulación Física, Cognitiva y Afectiva de Mayores (ROSES), funded by the Ministerio de Ciencia, Innovación y Universidades and Universidad Carlos III de Madrid (Read & Publish Agreement CRUE-CSIC 2022).

Sustainable forest management is concerned with the management of forests according to the principles of sustainable development. As a contribution to the field, this paper combines the Vehicle Routing Problem (VRP) (in which the vehicles are harvesters) with the Multiple Stock Size Cutting Stock Problem under uncertainty (in which the stock is logs). We present an Integer Linear Program that dynamically combines the cutting of the uncertain stock with vehicle routing, and uses it to address real-life problems. In experiments on real data from the forestry harvesting industry, we show that it outperforms a commonly used metaheuristic algorithm, This publication has emanated from research conducted with the fnancial support of Science Foundation Ireland under Grant Number 12/RC/2289-P2 which is co-funded under the European Regional Development Fund. We would also like to acknowledge the support of the Science Foundation Ireland CONFIRM Centre for Smart Manufacturing, Research Code 16/RC/3918

In software engineering, observability is the ability to determine the current state of a software system based on its external outputs or signals such as metrics, logs, or traces. Web engineers rely on the web browser console as the primary tool to monitor the client-side of web applications during end-to-end tests. However, this is a manual and time-consuming task due to the different browsers available. This paper presents BrowserWatcher, an open-source browser extension providing cross-browser capabilities to observe web applications and automatically gather browser console logs in different browsers (e.g., Chrome, Firefox, or Edge). We have leveraged this extension to conduct an empirical study analyzing the browser console of the top-50 public websites manually and automatically. The results show that BrowserWatcher gathers all the well-known log categories such as console or error traces. It also reveals that each web browser additionally includes other types of logs, which differ among browsers, thus providing distinct pieces of information for the same website. This work was partially supported in part by the Ministerio de Ciencia e Innovación-Agencia Estatal de Investigación, Spain (10.13039/501100011033) through the H2O Learn project under Grant PID2020-112584RB-C31, in part by the Madrid Regional Government through the e-Madrid-CM Project, Spain under Grant S2018/TCS-4307, and in part supported by the Comunidad de Madrid and Universidad Politécnica de Madrid, Spain through the V-PRICIT Research Programme Apoyo a la realización de Proyectos de I+D para jóvenes investigadores UPM-CAM, under Grant APOYOJOVENES-QINIM8-72-PKGQ0J. Funding for Article Processing Charge (APC): Universidad Carlos III de Madrid (Read & Publish Agreement CRUE-CSIC 2023).

Producción Científica, Heterogeneous systems with several kinds of devices, such as multi-core CPUs, GPUs, FPGAs, among others, are now commonplace. Exploiting all these devices with device-oriented programming models, such as CUDA or OpenCL, requires expertise and knowledge about the underlying hardware to tailor the application to each specific device, thus degrading performance portability. Higher-level proposals simplify the programming of these devices, but their current implementations do not have an efficient support to solve problems that include frequent bursts of computation and communication, or input/output operations. In this work we present CtrlEvents, a new heterogeneous runtime solution which automatically overlaps computation and communication whenever possible, simplifying and improving the efficiency of data-dependency analysis and the coordination of both device computations and host tasks that include generic I/O operations. Our solution outperforms other state-of-the-art implementations for most situations, presenting a good balance between portability, programmability and efficiency., Ministerio de Ciencia e Innovación - FEDER (TIN2017-88614-R), Junta de Castilla y León (VA226P20), Ministerio de Ciencia e Innovación - AEI and European Union NextGenerationEU/PRTR (TED2021–130367B–I00 and MCIN/AEI/10.13039/501100011033)

In robotic systems, both software and hardware components are equally important. However, scant attention has been devoted until now in order to detect anomalies/failures affecting the software component of robots while many proposals exist aimed at detecting physical anomalies. To bridge this gap, the present paper focuses on the study of anomalies affecting the software performance of a robot by using a novel visualization tool. Unsupervised visualization methods from the machine learning field are applied in order to upgrade the recently proposed Hybrid Unsupervised Exploratory Plots (HUEPs). Furthermore, Curvilinear Component Analysis and t-distributed stochastic neighbor embedding are added to the original HUEPs formulation and comprehensively compared. Furthermore, all the different combinations of HUEPs are validated in a real-life scenario. Thanks to this intelligent visualization of robot status, interesting conclusions can be obtained to improve anomaly detection in robot performance., Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature.

La presente investigación realizó el desarrollo de un software denominado Hope basado en realidad aumentada (RA) y enfocado en reforzar procesos de enseñanza aprendizaje de danza para niños con Trastorno de Espectro Autista (TEA). Para la consecución de este trabajo se realizó un levantamiento de datos recogiendo criterios y experiencias de un grupo multidisciplinario, 15 personas participaron en el estudio escogidos mediante una muestra aleatoria simple, mediante herramientas como encuesta y entrevista. A partir de ello se verificó algunos aspectos a tomar en cuenta previo al análisis, diseño, desarrollo e implementación del producto, tomando en cuenta la ISO 25010 referente a buscar la calidad en el software y la TR 29138-1 enfocadas en las necesidades de accesibilidad en personas con algún tipo de discapacidad. Se evidencia la existencia de características particularidades que deben ser definidas de acuerdo con el propósito, enfoque y los objetivos que persigue el software, se armonizan criterios técnicos, psicológicos, pedagógicos. Entre los resultados más destacados se observa que se debe tomar en cuenta criterios de accesibilidad y de usabilidad, la propuesta se basa en la creación de un espacio lúdico el mismo que causa una afinidad especial en este tipo de usuarios., The present investigation carried out the development of a software called Hope based on augmented reality (AR) and focused on reinforcing dance teaching-learning processes for children with Autism Spectrum Disorder (ASD). To achieve this work, a data survey was carried out, collecting criteria and experiences of a multidisciplinary group, 15 people participated in the study chosen through a simple random sample, using tools such as survey and interview. From this, some aspects to be taken into account prior to the analysis, design, development and implementation of the product were verified, taking into account ISO 25010 referring to seeking quality in software and TR 29138-1 focused on accessibility needs. in people with some type of disability. The existence of particular characteristics that must be defined according to the purpose, approach and objectives pursued by the software is evident, technical, psychological, pedagogical criteria are harmonized. Among the most outstanding results, it is observed that accessibility and usability criteria must be taken into account, the proposal is based on the creation of a playful space that causes a special affinity in this type of user., Laboratorio de Investigación en Nuevas Tecnologías Informáticas

In the last years, our society's high energy demand has led to the proposal of novel ways of consuming and producing electricity. In this sense, many countries have encouraged micro generation, including the use of renewable sources such as solar irradiation and wind generation, or considering the insertion of electric vehicles as dispatchable units on the grid. This work addresses the Optimal active&-reactive power dispatch (OARPD) problem (a type of optimal power flow (OPF) task) in microgrids considering electric vehicles. We used the modified IEEE 57 and IEEE 118 bus-systems test scenarios, in which thermoelectric generators were replaced by renewable generators. In particular, under the IEEE 118 bus system, electric vehicles were integrated into the grid. To solve the OARDP problem, we proposed the use and improvement of the Canonical Differential Evolutionary Particle Swarm Optimization (C-DEEPSO) algorithm. For further refinement in the search space, C-DEEPSO relies on local search operators. The results indicated that the proposed improved C-DEEPSO was able to show generation savings (in terms ofmillions of dollars) acting as a dispatch controller against two algorithms based on swarm intelligence., European Commission, Agencia Estatal de Investigación, Comunidad de Madrid

This survey reviews the scientific literature on techniques for reducing interference in real-time multicore systems, focusing on the approaches proposed between 2015 and 2020. It also presents proposals that use interference reduction techniques without considering the predictability issue. The survey highlights interference sources and categorizes proposals from the perspective of the shared resource. It covers techniques for reducing contentions in main memory, cache memory, a memory bus, and the integration of interference effects into schedulability analysis. Every section contains an overview of each proposal and an assessment of its advantages and disadvantages. This work was supported in part by the Comunidad de Madrid Government "Nuevas Técnicas de Desarrollo de Software de Tiempo Real Embarcado Para Plataformas. MPSoC de Próxima Generación" under Grant IND2019/TIC-17261.

Background Bodily illusions can be used to investigate the experience of being in a body by manipulating the underlying processes of multisensory integration. Research suggests that people with eating disorders (EDs) may have impairments in visual, interoceptive, proprioceptive, and tactile bodily perception. Furthermore, people with EDs also show abnormalities in integrating multisensory visuo-tactile and visual-auditory signals related to the body, which may contribute to the development of body image disturbances. Visuo-auditory integration abnormalities have been observed also in people with subthreshold ED symptomatology. However, it remains unclear whether these impairments are specific to bodily signals or if they extend to any auditory signals. Methods We will recruit 50 participants (aged 18–24; females assigned at birth) with ED symptomatology (subthreshold group) and 50 control participants. The Eating Disorder Examination Questionnaire will be administered to screen for ED symptomatology and divide the sample into two groups accordingly (control and subthreshold group using a clinical cut-off score of 2.8). The strength of both illusions will be measured implicitly with estimations of body part position and size, and explicitly with self-report questionnaires. As a secondary aim, regression analysis will be run to test the predictive role of susceptibility for both illusions on interoceptive body awareness (measured by the Multidimensional Assessment of Interoceptive Awareness Scale) and sensory-processing sensitivity (measured by the Highly Sensitive Person Scale). Discussion Our study may contribute to our understanding of the mechanisms underlying body image disturbances. The results may pave the way for novel clinical interventions targeting early symptoms prior to the development of the disorder in young females.

In recent years, artificial intelligence has played an important role in education, wherein one of the most commonly used applications is forecasting students’ academic performance based on personal information such as social status, income, address, etc. This study proposes and develops an artificial neural network model capable of determining whether a student will pass a certain class without using personal or sensitive information that may compromise student privacy. For model training, we used information regarding 32,000 students collected from The Open University of the United Kingdom, such as number of times they took the course, average number of evaluations, course pass rate, average use of virtual materials per date and number of clicks in virtual classrooms. Attributes selected for the model are as follows: 93.81% accuracy, 94.15% precision, 95.13% recall, and 94.64% F1-score. These results will help the student authorities to take measures to avoid withdrawal and underachievement.

Resumos das teses defendidas e homologadas no Programa de Pós-Graduação em Informática na Educação entre os meses de janeiro e abril de 2020.

In this review essay, we give a detailed synopsis of the twelve contributions which are collected in a Special Issue in Frontiers Ecology and Evolution, based on the research topic “Current Thoughts on the Brain-Computer Analogy—All Metaphors Are Wrong, But Some Are Useful.” The synopsis is complemented by a graphical summary, a matrix which links articles to selected concepts. As first identified by Turing, all authors in this Special Issue recognize semantics as a crucial concern in the brain-computer analogy debate, and consequently address a number of such issues. What is missing, we believe, is the distinction between metaphor and analogy, which we reevaluate, describe in some detail, and offer a definition for the latter. To enrich the debate, we also deem necessary to develop on the evolutionary theories of the brain, of which we provide an overview. This article closes with thoughts on creativity in Science, for we concur with the stance that metaphors and analogies, and their esthetic impact, are essential to the creative process, be it in Sciences as well as in Arts.

This paper discusses a prominent way in which social simulations can contribute (and have contributed) to the advance of science; namely, by refuting some of our incorrect beliefs about how the real world works. More precisely, social simulations can produce counter-examples that reveal something is wrong in a prevailing scientific assumption. Indeed, here we argue that this is a role that many well-known social simulation models have played, and it may be one of the main reasons why such well-known models have become so popular. To test this hypothesis, here we examine several popular models in the social simulation literature and we find that all these models are most naturally interpreted as providers of compelling and reproducible (computer-generated) evidence that refuted some assumption or belief in a prevailing theory. By refuting prevailing theories, these models have greatly advanced science and, in some cases, have even opened a new field of research.

In recent years, the scientific community has shown a special interest in the research, deployment, and use of computational, heterogeneous, and autonomous entities as intelligent agents to work together for different emerging or critical tasks in the field. The literature proposes and evidences the benefits of providing proactive organisational capability to these entities in dynamic scenarios. The motivation is the following: all decisions that can be planned by these entities as agents are no longer delegated to the users, so that the automation of tasks lightens the workload and frustration from users, with the goal of maximising the teams’ effectiveness. AI deployment through Multi-Agent Systems requires modelling that proactivity in a changing scenario. Additionally, ensuring the achievement of all objectives is even more challenging when considering events and risks that compromise the integrity of the team. However, these scenarios require distributed solutions with capabilities beyond traditional and centralised fault-tolerant techniques. Resilient AI research proposes systems that can recover as gracefully and efficiently as possible from unavoidable and high-impact circumstances. The deployment of resilient systems implies the incremental implementation of the different principles and the different synergies that compose it, involving a challenge in defining a generalised model of the resilient system. The question that arises is how Resilience can be characterised through a distributed system in a dynamic scenario in such a way that it would provide an advantage in the search, guarantee and quality of the solution. This Ph.D. thesis aims to answer this question by proposing a first approach to solving the resilient Team Formation problem. This constrained-based problem consists of a well-known mathematical framework to represent a dynamic scenario as a set of agents, variables, and restrictions to be satisfied. Thus, the goal in its resilient version is to assemble a team of agents capable to perform a set of tasks with the minimum cost and guaranteeing its solution in the event of a possible failure or deprivation of a number of agents. For this purpose, it is proposed to introduce the principle of Stabilisation by characterising the agent communication when forming the team. Thus, a resilient solution now includes the search for a team that also meets and preserves certain conditions in its graph structure. As contributions of this work, an extended generalisation of the resilient Team Formation problem that includes the representation of the communication between agents is presented, as well as the design and implementation of an optimal and efficient distributed algorithm that solves this problem. The results support that stabilisation is a valid component within a resilient system that allows the characterisation of both distributed and intelligent nature required in these scenarios. RESUMEN En los últimos años, la comunidad cientifica ha mostrado un especial interés en la investigación y desarrollo de entidades computacionales, heterogéneas y autónomas para trabajar en conjunto y de forma colaborativa, en forma de agentes inteligentes, en la realización de diferentes tareas emergentes o críticas sobre el terreno. La literatura propone y evidencia los beneficios de dotar de una capacidad de organización proactiva a estas entidades en escenarios dinámicos. La motivación del presente trabajo es la siguiente: las decisiones que pueden ser planificadas y automatizadas por un conjunto de agentes inteligentes no necesitan ser delegadas a los usuarios, con el objetivo de maximizar la efectividad y el rendimiento del equipo y sus tareas. Desplegar una inteligencia artificial sobre estos sistemas multi-agente requiere caracterizar esa proactividad en un escenario cambiante. Además, garantizar la consecución de todos los objetivos supone un reto aún mayor si consideramos los eventos y riesgos que comprometen la integridad del equipo. Tradicionalmente, el enfoque en estos escenarios pasa por soluciones parcialmente centralizadas junto con el uso de heurísticas basadas en técnicas de tolerancia a fallos. Sin embargo, estos escenarios requieren soluciones distribuidas con la capacidad adicional de poder recuperarse de la forma más grácil y eficiente posible ante circunstancias inevitables y de gran impacto, lo que se conoce en la literatura como sistemas resilientes. La aplicación de la Inteligencia Artificial Resiliente se basa en la implementación incremental de los diferentes principios y conceptos que la componen. Sin embargo, las sinergias entre los distintos componentes hacen complicado un diseño generalizado y que incluya una solución óptima al mismo tiempo. Surge entonces la pregunta de cómo la resiliencia puede ser caracterizada a través de un sistema distribuido, de tal forma que modelar un problema representativo de estos escenarios dinámicos suponga una ventaja en la búsqueda, garantía, y calidad de la solución. La presente tesis doctoral pretende contestar a esta pregunta proponiendo una primera aproximación que resuelva el problema conocido de formación de equipos resiliente. Este problema distribuido está basado en restricciones, un marco matemático para representar un escenario dinámico como un conjunto de agentes asignando variables sujetas a satisfacer una serie de restricciones o limitaciones. El objetivo del problema en su versión resiliente consiste en reunir un equipo de agentes para realizar un conjunto de tareas con el mínimo coste y garantizando su solución ante una posible avería o eliminación de un número de agentes. Para ello, se propone incorporar al problema la caracterización de la comunicación de los agentes a la hora de conformar el equipo. Se extiende así la resiliencia introduciendo por primera vez el principio de la estabilización. Esta definición establece en el problema la búsqueda de un equipo que además cumpla y preserve ante cualquier circunstancia ciertas condiciones en su estructura basándose en ciertas propiedades bien conocidas de la teoría de grafos, como su conectividad o densidad. Como contribuciones más importantes de este trabajo, por un lado, se presenta una generalización del problema de formación de equipos resiliente que incluye la representación de la comunicación entre agentes, así como el diseño e implementación de un algoritmo distribuido que resuelve de forma parametrizada, óptima y eficiente este problema. Por otro lado, se analizan y discuten las diferentes implicaciones de la estabilización del equipo frente a aspectos resilientes como su robustez y recuperabilidad, así como su eficiencia y escalabilidad. Como resultado, se demuestra que la estabilización es un componente válido dentro de un sistema resiliente que permite caracterizar la naturaleza distribuida e inteligente requerida en estos escenarios.

The use of videoconferencing is on the rise after COVID-19, being common to look at the screen and see someone typing. A side-channel attack may be launched to infer the text written from the face image. In this paper, we analyse the feasibility of such an attack, being the first proposal which work with a complete keyset (50 keys) and natural texts. We use different scenarios, lighting conditions and natural texts to increase realism. Our study involves 30 participants, who typed 49,365 keystrokes. We characterize the effect of lighting, gender, age and use of glasses. Our results show that on average 13.71% of keystrokes are revealed without error, and up to 31.8%, 52.5% and 61.2% are guessed with a maximum error of 1, 2 and 3 keys, respectively. This work was supported by the Madrid Goverment (Comunidad de Madrid-Spain) under the multianual agreement with UC3M (“fostering young doctor research”, DEPROFAKECM-UC3M) and in the context of the V PRICIT research and technological innovation regional program; by CAM by grant CYNAMON P2018/TCS-4566-CM, cofunded with ERDF; and by Min. of Science and Innovation of Spain by grant ODIO PID2019-111429RB-C21 (AEI/10.13039/501100011033).

This is the peer reviewed version of the following article: Diez, P. [et al.]. Nonlinear dimensionality reduction for parametric problems: a kernel proper orthogonal decomposition. "International journal for numerical methods in engineering", 30 Desembre 2021, vol. 122, núm. 24, p. 7306-7327, which has been published in final form at DOI: 10.1002/nme.6831. This article may be used for non-commercial purposes in accordance with Wiley Terms and Conditions for Self-Archiving. Reduced-order models are essential tools to deal with parametric problems in the context of optimization, uncertainty quantification, or control and inverse problems. The set of parametric solutions lies in a low-dimensional manifold (with dimension equal to the number of independent parameters) embedded in a large-dimensional space (dimension equal to the number of degrees of freedom of the full-order discrete model). A posteriori model reduction is based on constructing a basis from a family of snapshots (solutions of the full-order model computed offline), and then use this new basis to solve the subsequent instances online. Proper orthogonal decomposition (POD) reduces the problem into a linear subspace of lower dimension, eliminating redundancies in the family of snapshots. The strategy proposed here is to use a nonlinear dimensionality reduction technique, namely, the kernel principal component analysis (kPCA), in order to find a nonlinear manifold, with an expected much lower dimension, and to solve the problem in this low-dimensional manifold. Guided by this paradigm, the methodology devised here introduces different novel ideas, namely, 1) characterizing the nonlinear manifold using local tangent spaces, where the reduced-order problem is linear and based on the neighboring snapshots, 2) the approximation space is enriched with the cross-products of the snapshots, introducing a quadratic description, 3) the kernel for kPCA is defined ad hoc, based on physical considerations, and 4) the iterations in the reduced-dimensional space are performed using an algorithm based on a Delaunay tessellation of the cloud of snapshots in the reduced space. The resulting computational strategy is performing outstandingly in the numerical tests, alleviating many of the problems associated with POD and improving the numerical accuracy. Generalitat de Catalunya, 2017-SGR-1278; Ministerio de Ciencia e Innovación, CEX2018-000797-S; PID2020-113463RB-C32; PID2020-113463RB-C33

The World Wide Web has become a common platform for interactive software development. Most web applications feature custom user interfaces used by millions of people every day. Information architecture addresses the structural design of information to build quality web applications with improved usability of content, navigation, and findability. One of the most frequently utilized information architecture methods is card sorting—an affordable, user-centered approach for eliciting and evaluating categories and navigable items. Card sorting facilitates decision-making during the development process based on users’ mental models of a given application domain. However, although the qualitative analysis of card sorts has become common practice in information architecture, the quantitative analysis of card sorting is less widely applied. The reason for this gap is that quantitative analysis often requires the use of customized techniques to extract meaningful information for decision-making. To facilitate this process and support the structuring of information, we propose a methodology for the quantitative analysis of card-sorting results in this paper. The suggested approach can be systematically applied to provide clues and support for decisions. These might significantly impact the design and, thus, the final quality of the web application. Therefore, the approach includes proper goodness values that enable comparisons among the results of the methods and techniques used and ensure the suitability of the analyses performed. Two publicly available datasets were used to demonstrate the key issues related to the interpretation of card sorting results and the overall suitability and validity of the proposed methodology, This work was partially supported by the Spanish Government [grant number RTI2018-095255-B-I00]; and the Madrid Research Council [Grant Number P2018/TCS-4314]

Quite often a machine learning problem lends itself to be split in several well-defined subproblems, or tasks. The goal of Multi-Task Learning (MTL) is to leverage the joint learning of the problem from two different perspectives: on the one hand, a single, overall model, and on the other hand task-specific models. In this way, the found solution by MTL may be better than those of either the common or the task-specific models. Starting with the work of Evgeniou et al., support vector machines (SVMs) have lent themselves naturally to this approach. This paper proposes a convex formulation of MTL for the L1-, L2- and LS-SVM models that results in dual problems quite similar to the single-task ones, but with multi-task kernels; in turn, this makes possible to train the convex MTL models using standard solvers. As an alternative approach, the direct optimal combination of the already trained common and task-specific models can also be considered. In this paper, a procedure to compute the optimal combining parameter with respect to four different error functions is derived. As shown experimentally, the proposed convex MTL approach performs generally better than the alternative optimal convex combination, and both of them are better than the straight use of either common or task-specific models, With partial support from Spain’s grant TIN2016-76406-P. Work supported also by the UAM–ADIC Chair for Data Science and Machine Learning.

Machine learning (ML) techniques such as classifiers are used in many applications, some of which are related to safety or critical systems. In this case, correct processing is a strict requirement and thus ML algorithms (such as for classification) must be error tolerant. A naive approach to implement error tolerant classifiers is to resort to general protection techniques such as modular redundancy. However, modular redundancy incurs in large overheads in many metrics such as hardware utilization and power consumption that may not be acceptable in applications that run on embedded or battery powered systems. Another option is to exploit the algorithmic properties of the classifier to provide protection and error tolerance at a lower cost. This paper explores this approach for a widely used classifier, the k Nearest Neighbors ( k NNs), and proposes an efficient scheme to protect it against errors. The proposed technique is based on a time-based modular redundancy (TBMR) scheme. The proposed scheme exploits the intrinsic redundancy of k NNs to drastically reduce the number of re-computations needed to detect errors. This is achieved by noting that when voting among the k nearest neighbors has a large majority, an error in one of the voters cannot change the result, hence voting margin (VM). This observation has been refined and extended in the proposed VM scheme to also avoid re-computations in some cases in which the majority vote is tight. The VM scheme has been implemented and evaluated with publicly available data sets that cover a wide range of applications and settings. The results show that by exploiting the intrinsic redundancy of the classifier, the proposed scheme is able to reduce the cost compared to modular redundancy by more than 60 percent in all configurations evaluated. Pedro Reviriego and Josée Alberto Hernández would like to acknowledge the support of the TEXEO project TEC2016-80339-R funded by the Spanish Ministry of Economy and Competitivity and of the Madrid Community research project TAPIR-CM Grant no. P2018/TCS-4496.

A well-established approach to reasoning about loops during program analysis is to capture the effect of a loop by extracting recurrences from the loop; these express relationships between the values of variables, or program properties such as cost, on successive loop iterations. Recurrence solvers are capable of computing closed forms for some recurrences, thus deriving precise relationships capturing the complete loop execution. However, many recurrences extracted from loops cannot be solved, due to their having multiple recursive cases or multiple arguments. In the literature, several techniques for approximating the solution of unsolvable recurrences have been proposed. The approach presented in this paper is to define transformations based on regular path expressions and loop counters that (i) transform multi-path loops to single-path loops, giving rise to recurrences with a single recursive case, and (ii) transform multi-argument recurrences to single-argument recurrences, thus enabling the use of recurrence solvers on the transformed recurrences. Using this approach, precise solutions can sometimes be obtained that are not obtained by approximation methods., In Proceedings HCVS 2021, arXiv:2109.03988

The multidimensional classification of multivariate timeseries deals with the assignment of multiple classes to time‐ordered data described by a set of feature variables. Although this challenging task has received almost no attention in the literature, it is present in a wide variety of domains, such as medicine, finance or industry. The complexity of this problem lies in two nontrivial tasks, the learning with multivariate time series in continuous time and the simultaneous classification of multiple class variables that may show dependencies between them. These can be addressed with different strategies, but most of them may involve a difficult preprocessing of the data, high space and classification complexity or ignoring useful interclass dependencies. Additionally, no attention has been given to the development of new multidimensional classifiers of time series based on probabilistic graphical models, even though transparent models can facilitate further understanding of the domain. In this paper, a novel probabilistic graphical model is proposed, which is able to classify a discrete multivariate temporal sequence into multiple class variables while modeling their dependencies. This model extends continuous time Bayesian networks to the multidimensional classification problem, which are able to explicitly represent the behavior of time series that evolve over continuous time. Different methods for the learning of the parameters and structure of the model are presented, and numerical experiments on synthetic and real‐world data show encouraging results in terms of performance and learning time with respect to independent classifiers, the current alternative approach under the continuous time Bayesian network paradigm.

La ciberseguridad ha destacado en los ultimos anos con el objetivo de proteger los sistemas de informacion. Se han utilizado diversos metodos, tecnicas y herramientas para obtener el mayor provecho de las vulnerabilidades existentes de estos sistemas. Por lo tanto, es imprescindible el desarrollo y mejora en nuevas tecnologias, asi como de los sistemas de deteccion de intrusiones que permiten detectar posibles amenazas. Sin embargo, el uso de estas tecnologias requiere personal altamente cualificado en materia de ciberseguridad para analizar los resultados y mejorar la accuracy en sus resultados. Por lo tanto, esto genera la necesidad de investigar y desarrollar nuevos sistemas de ciberseguridad con altas prestaciones que permitan analizar y resolver de manera eficiente dichos resultados. Esta investigacion presenta la aplicacion de tecnicas de aprendizaje automatico para clasificacion de trafico real, con el objetivo de identificar posibles ataques. El estudio se ha realizado utilizando herramientas de aprendizaje automatico aplicando algoritmos de tipo deep learning como multi layer perceptron y long short term memory. Adicionalmente, este documento presenta una comparacion entre los resultados obtenidos por la aplicacion de los algoritmos anteriormente mencionados y algoritmos que no son de tipo deep learning como: random forest y decision tree. Finalmente, se exponen los resultados obtenidos demostrando que el algoritmo long short term memory es el que brinda los mejores resultados en relacion de accuracy y perdida logaritmica. Palabras clave: Inteligencia Artificial, LSTM, redes neuronales, sistema de deteccion de intrusiones, UGR16