Search

This study aims to understand the development trends and research structure of articles on artificial intelligence (AI) and information processing in the past 10 years. In particular, this study analyzed 13,294 papers published from 2012 to 2021 in the Web of Science, used the bibliometric analysis method to visualize the data of the papers, and drew a scientific knowledge map. By exploring the development of mainstream journals, author and country rankings, keyword evolution, and research field rankings in the past 10 years, this study uncovered key trends affecting AI progress and information processing that provide insights and serve as an important reference for future AI research and information processing. The results revealed a gradual increase in publications over the past decade, with explosive growth after 2020. The most prolific researchers in this field were Xu, Z.S.; Pedrycz, W.; Herrera-Viedma, E.; the major contributing countries were China, the USA, and Spain. In the AI and information processing research, keywords including "Deep learning", "Machine learning", and "Feature extraction" are components that play a crucial role. Additionally, the most representative research areas were "Engineering", "Operations Research and Management Science", and "Automation Control Systems". Overall, this study used bibliometric analysis to provide an overview of the latest trends in artificial intelligence and information processing. Although AI and information processing have been applied to various research areas, many other sub-topics can be further applied. Based on the findings, this study presented research insights and proposed suggestions for future research directions on AI and information processing. [ABSTRACT FROM AUTHOR]

Respiratory viruses, such as COVID-19, are spread over time and space based on human-to-human interactions. Human mobility plays a key role in the propagation of the virus. Different types of sensors in smart cities are able to continuously monitor traffic-related human mobility, showing the impact of COVID-19 on traffic volumes and patterns. In a similar way, traffic volumes measured by smart traffic sensors provide a proxy variable to capture human mobility, which is expected to have an impact on new COVID-19 infections. Adding traffic data from smart city sensors to machine learning models designed to estimate upcoming COVID-19 incidence values should provide optimized results compared to models based on COVID-19 data alone. This paper proposes a novel model to extract spatio-temporal patterns in the spread of the COVID-19 virus for short-term predictions by organizing COVID-19 incidence and traffic data as interrelated temporal sequences of spatial images. The model is trained and validated with real data from the city of Madrid in Spain for 84 weeks, combining information from 4372 traffic measuring points and 143 COVID-19 PCR test centers. The results are compared with a baseline model designed for the extraction of spatio-temporal patterns from COVID-19-only sequences of images, showing that using traffic information enhances the results when forecasting a new wave of infections (MSE values are reduced by a 70% factor). The information that traffic data has on the spread of the COVID-19 virus is also analyzed, showing that traffic data alone is not sufficient for accurate COVID-19 forecasting. [ABSTRACT FROM AUTHOR]

The aim of this work is to develop and evaluate a deep classifier that can effectively prioritize Emergency Medical Call Incidents (EMCI) according to their life-threatening level under the presence of dataset shifts. We utilized a dataset consisting of 1982746 independent EMCI instances obtained from the Health Services Department of the Region of Valencia (Spain), with a time span from 2009 to 2019 (excluding 2013). The dataset includes free text dispatcher observations recorded during the call, as well as a binary variable indicating whether the event was life-threatening. To evaluate the presence of dataset shifts, we examined prior probability shifts, covariate shifts, and concept shifts. Subsequently, we designed and implemented four deep Continual Learning (CL) strategies-cumulative learning, continual fine-tuning, experience replay, and synaptic intelligence-alongside three deep CL baselines-joint training, static approach, and single fine-tuning-based on DistilBERT models. Our results demonstrated evidence of prior probability shifts, covariate shifts, and concept shifts in the data. Applying CL techniques had a statistically significant (α=0.05) positive impact on both backward and forward knowledge transfer, as measured by the F1-score, compared to non-continual approaches. We can argue that the utilization of CL techniques in the context of EMCI is effective in adapting deep learning classifiers to changes in data distributions, thereby maintaining the stability of model performance over time. To our knowledge, this study represents the first exploration of a CL approach using real EMCI data., Competing Interests: Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2024 The Author(s). Published by Elsevier Ltd.. All rights reserved.)

In this paper, we leverage multimodal data to classify minerals using a multi-stream neural network. In a previous study on the Tinto dataset, which consisted of a 3D hyperspectral point cloud from the open-pit mine Corta Atalaya in Spain, we successfully identified mineral classes by employing various deep learning models. However, this prior work solely relied on hyperspectral data as input for the deep learning models. In this study, we aim to enhance accuracy by incorporating multimodal data, which includes hyperspectral images, RGB images, and a 3D point cloud. To achieve this, we have adopted a graph-based neural network, known for its efficiency in aggregating local information, based on our past observations where it consistently performed well across different hyperspectral sensors. Subsequently, we constructed a multi-stream neural network tailored to handle multimodality. Additionally, we employed a channel attention module on the hyperspectral stream to fully exploit the spectral information within the hyperspectral data. Through the integration of multimodal data and a multi-stream neural network, we achieved a notable improvement in mineral classification accuracy: 19.2%, 4.4%, and 5.6% on the LWIR, SWIR, and VNIR datasets, respectively. [ABSTRACT FROM AUTHOR]

Airborne pollen can trigger allergic rhinitis and other respiratory diseases in the synthesised population, which makes it one of the most relevant biological contaminants. Therefore, implementing accurate forecast systems is a priority for public health. The current forecast models are generally useful, but they falter when long time series of data are managed. The emergence of new computational techniques such as the LSTM algorithms could constitute a significant improvement for the pollen risk assessment. In this study, several LSTM variants were applied to forecast monthly pollen integrals in Málaga (southern Spain) using meteorological variables as predictors. Olea and Urticaceae pollen types were modelled as proxies of different annual pollen curves, using data from the period 1992-2022. The aims of this study were to determine the LSTM variants with the highest accuracy when forecasting monthly pollen integrals as well as to compare their performance with the traditional pollen forecast methods. The results showed that the CNN-LSTM were the most accurate when forecasting the monthly pollen integrals for both pollen types. Moreover, the traditional forecast methods were outperformed by all the LSTM variants. These findings highlight the importance of implementing LSTM models in pollen forecasting for public health and research applications., Competing Interests: Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2023 Elsevier Ltd. All rights reserved.)

Energy management systems allow the Smart Grids industry to track, improve, and regulate energy use. Particularly, demand-side management is regarded as a crucial component of the entire Smart Grids system. Therefore, by aligning utility offers with customer demand, anticipating future energy demands is essential for regulating consumption. An updated examination of several forecasting techniques for projecting energy short-term load forecasts is provided in this article. Each class of algorithms, including statistical techniques, Machine Learning, Deep Learning, and hybrid combinations, are comparatively evaluated and critically analyzed, based on three real consumption datasets from Spain, Germany, and the United States of America. To increase the size of tiny training datasets, this paper also proposes a data augmentation technique based on Generative Adversarial Networks. The results show that the Deep Learning-hybrid model is more accurate than traditional statistical methods and basic Machine Learning procedures. In the same direction, it is demonstrated that more comprehensive datasets assisted by complementary data, such as energy generation and weather, may significantly boost the accuracy of the models. Additionally, it is also demonstrated that Generative Adversarial Networks-based data augmentation may greatly improve algorithm accuracy. [ABSTRACT FROM AUTHOR]

In this paper, we design an Anomaly Detection (AD) framework for mobile data traffic, capable of identifying different types of anomalous events generated by flash crowds in metropolitan areas. We state the problem using a semi-supervised approach and exploit the great performance of different Recurrent Neural Network (RNN) models to learn the temporal context of input sequences. Our proposal processes real traffic traces from the unencrypted LTE Physical Downlink Control Channel (PDCCH) of an operative network, gathered during an extensive measurement campaign in two major cities in Spain. The AD framework is designed to perform: i) a-posteriori analysis to understand users’ behavior and urban environment variations; ii) real-time analysis to automatically and on-the-fly alert urban anomalies; and iii) estimation of the duration of the periods identified as anomalous. Numerical results show the higher performance of our AD framework compared to classic AD algorithms and confirm that the proposed framework predicts anomalous behaviours with high accuracy and regardless of their cause. [ABSTRACT FROM AUTHOR]

In this paper, we propose a twofold methodology for visual detection and recognition of different types of city dumpsters, with minimal human labeling of the image data set. Firstly, we carry out transfer learning by using Google Inception-v3 convolutional neural network, which is retrained with only a small subset of labeled images out of the whole data set. This first classifier is then improved with a semi-supervised learning based on retraining for two more rounds, each one increasing the number of labeled images but without human supervision. We compare our approach against both to a baseline case, with no incremental retraining, and the best case, assuming we had a fully labeled data set. We use a data set of 27,624 labeled images of dumpsters provided by Ecoembes, a Spanish nonprofit organization that cares for the environment through recycling and the eco-design of packaging in Spain. Such a data set presents a number of challenges. As in other outdoor visual tasks, there are occluding objects such as vehicles, pedestrians and street furniture, as well as other dumpsters whenever they are placed in groups. In addition, dumpsters have different degrees of deterioration which may affect their shape and color. Finally, 35% of the images are classified according to the capacity of the container, which contains a feature which is hard to assess in a snapshot. Since the data set is fully labeled, we can compare our approach both against a baseline case, doing only the transfer learning using a minimal set of labeled images, and against the best case, using all the labels. The experiments show that the proposed system provides an accuracy of 88%, whereas in the best case it is 93%. In other words, the method proposed attains 94% of the best performance. [ABSTRACT FROM AUTHOR]

Time series forecasting of meteorological variables, such as the hourly air temperature, has multiple benefits for industry, agriculture, and the environment. Due to the high accuracy required for the associated short-term predictions, traditional methods cannot satisfy the requirements and generally ignore spatial dependencies. This paper proposes a deep Graph Convolutional Long Short Term Memory Neural Network (GCN-LSTM) technique to tackle the time series prediction problem in air temperature forecasting. In the proposed methodology, temporal and spatial-based imputation approaches have been employed to recover the weather variables missing values. The proposed approach is validated using real, open weather data from 37 meteorological stations in Spain. Performed analysis indicates that GCN-LSTM showed superior performance when compared with various state-of-the-art Deep Learning based models found in the literature, resulting in a more robust and computationally efficient model for forecasting air temperature in many meteorological stations simultaneously. [ABSTRACT FROM AUTHOR]

In this paper, we propose different explicable forecasting approaches, based on inductive and evolutionary decision rules, for extreme low-visibility events prediction. Explicability of the processes given by the rules is in the core of the proposal. We propose two different methodologies: first, we apply the PRIM algorithm and evolution to obtain induced and evolved rules, and subsequently these rules and boxes of rules are used as a possible simpler alternative to ML/DL classifiers. Second, we propose to integrate the information provided by the induced/evolved rules in the ML/DL techniques, as extra inputs, in order to enrich the complex ML/DL models. Experiments in the prediction of extreme low-visibility events in Northern Spain due to orographic fog show the good performance of the proposed approaches. [ABSTRACT FROM AUTHOR]

Non-technical losses (NTL) in electricity utilities are responsible for major revenue losses. In this paper, we propose a novel end-to-end solution to self-learn the features for detecting anomalies and frauds in smart meters using a hybrid deep neural network. The network is fed with simple raw data, removing the need of handcrafted feature engineering. The proposed architecture consists of a long short-term memory network and a multi-layer perceptrons network. The first network analyses the raw daily energy consumption history whilst the second one integrates non-sequential data such as its contracted power or geographical information. The results show that the hybrid neural network significantly outperforms state-of-the-art classifiers as well as previous deep learning models used in NTL detection. The model has been trained and tested with real smart meter data of Endesa, the largest electricity utility in Spain. [ABSTRACT FROM AUTHOR]

Transfer learning methods reuse a deep learning model developed for a task on another task. Such methods have been remarkably successful in a wide range of image processing applications. Following the trend, few transfer learning based methods have been proposed for unsupervised multi-temporal image analysis and change detection (CD). Inspite of their success, the transfer learning based CD methods suffer from limited explainability. In this paper, we propose an explainable convolutional autoencoder model for CD. The model is trained in: 1) an unsupervised way using, as the bi-temporal images, patches extracted from the same geographic location; 2) a greedy fashion, one encoder and decoder layer pair at a time. A number of features relevant for CD is chosen from the encoder layer. To build an explainable model, only selected features from the encoder layer is retained and the rest is discarded. Following this, another encoder and decoder layer pair is added to the model in similar fashion until convergence. We further visualize the features to better interpret the learned features. We validated the proposed method on a Landsat-8 dataset obtained in Spain. Using a set of experiments, we demonstrate the explainability and effectiveness of the proposed model. [ABSTRACT FROM AUTHOR]

Deep learning techniques provide a novel framework for prediction and classification in decision-making procedures that are widely applied in different fields. Precision agriculture is one of these fields where the use of decision-making technologies provides better production with better costs and a greater benefit for farmers. This paper develops an intelligent framework based on a deep learning model for early prediction of crop frost to help farmers activate anti-frost techniques to save the crop. This model is based on a long short-term memory (LSTM) model and it is designed to predict low temperatures. The model is based on information from an IoT infrastructure deployed on two plots in Murcia (Southeast of Spain). Three experiments are performed; a cross validation to validate the model from the most pessimistic point of view, a validation of 24 consecutive hours of temperatures, in order to know 24 hours before the possible temperature drop and a comparison with two traditional time series prediction techniques, namely Auto Regressive Integrated Moving Average and the Gaussian process. The results obtained are satisfactory, being better the results of the LSTM, obtaining an average quadratic error of less than a Celsius degree and a determination coefficient R 2 greater than 0.95. [ABSTRACT FROM AUTHOR]

Extreme Wind Speed events (EWS) are responsible for the worst damages caused by wind in wind farms. An accurate estimation of the frequency and intensity of EWS is essential to avoid wind turbine damage and to minimize cut-out events in these facilities. In this paper we discuss how generative Data Augmentation (DA) techniques improve the performance of Machine Learning (ML) and Deep Learning (DL) algorithms in EWS prediction problems. These problems are usually tackled as classification tasks, which are highly unbalanced due to the small number of EWS events in wind farms. Different versions of Variational AutoEncoders (VAE) are proposed and analysed in this work (VAEs, Conditional VAEs (CVAEs) and Class-Informed VAEs (CI-VAE)) as generative DA techniques to balance the data in EWS problems, leading to better performance of the prediction systems. The proposed generative DA techniques have been compared against traditional DA algorithms in a real problem of EWS prediction in Spain, considering ERA5 reanalysis data as predictive variables. The results showed that the CI-VAE with a Convolutional Neural Network approach obtained the best results, with values of Precision 0.62, Recall 0.74 and F1 score 0.67, improving up to 4% the results of the method without data augmentation techniques. [Display omitted] • Generative data augmentation techniques can improve extreme wind prediction. • Variational Autoencoders modifications are used to generate synthetic wind extreme samples. • Experiments on real wind speed data from a Spanish wind farm are presented. • Comparisons with traditional data augmentation techniques show prediction improvements. [ABSTRACT FROM AUTHOR]

Purpose: This article draws upon evidence from a contemporary study of all-through schools (ATS) in three countries. ATS combine at least two stages of a child's education in a single establishment. Many admit children aged 3–19. Most children join the school at nursery or kindergarten level and continue there for their entire education before moving on to further or higher education. ATS are also called all-age, in some contexts, because they bring children of all ages together into the same school environment. Models of ATS vary internationally; hence, there is not one definition of an ATS. This article takes a comparative look at ATS in Iceland, Spain and Wales. The purpose of this article is to explore innovative pedagogies in ATS and to explore how far deeper learning occurs because of the integrated and inclusive model of schooling. The study focused on pedagogical practices in ATS and examined how far these innovative practices are considered by teachers to foster deeper learning outcomes. Design/methodology/approach: The article draws upon a three-year comparative research project that explored pedagogy, leadership and well-being in ATS. The article investigates pedagogy with a cross-cutting focus on enquiry and deeper learning from the perspective of leaders and teachers. Using focus groups and lesson observations, a qualitative case-study approach was utilised to gather evidence about the teaching and learning processes adopted in ATS. Semi-structured interviews were also conducted with school leaders. The analytical approach adopted was one of constant comparison with the prime aim of eliciting common themes across the data sets. In relation to the pedagogy theme and an exploration of pedagogical innovation, research questions included (1) How far do ATS foster innovative pedagogies?, (2) What are the leadership conditions that support innovative pedagogies? and (3) To what extent do innovative pedagogies promote deeper learning? Findings: Within and across the three education systems under investigation, the study found that all-through schooling engages students in a positive learning environment and provides innovative pedagogical processes associated with deeper learning. The article provides evidence about how deeper learning functions in ATS from different parts of the world and reflects on the way deeper learning is promoted by leaders and teachers, resulting in deeper learning for students. The evidence from this study reinforces that opportunities for pedagogical innovation and deeper learning within ATS occur because of flatter structures, more fluidity between different phases of learning and greater cross-over of teacher expertise. The study also highlights how leadership is a critical factor in creating the conditions for collective professional practices that foster pedagogical innovations to secure deeper learning. Findings suggest that leading for deeper learning is fundamentally concerned with creating the conditions for innovative learning environments that are equitable, inclusive, diverse and cross age ranges. Originality/value: Contemporary empirical studies of the deeper learning environments within ATS remain relatively rare; hence, this study provides new comparative and contemporary evidence that illuminates the nature of the pedagogical innovation and the leadership practices that support pedagogical innovation in these schools. It also highlights how professional collaboration and cross-phase working are at the heart of innovative pedagogies that support deeper learning. The study outlined in this article provides critical, new insights about pedagogical innovation and deeper learning within ATS settings. [ABSTRACT FROM AUTHOR]

Spain is the second country in Europe with the most swimming pools. However, the legal literature estimates that 20% of swimming pools are not declared or irregular.The administration has a corps of people who manually analyze satellite or drone images to detect illegal or irregular structures. This method is costly in terms of effort and time, and it is also a method based on the subjectivity of the person carrying it out. This proposal aims to design a platform that allows the automatic detection of irregular pools. Using geographic information tools (GIS) based on orthophotography, combined with advanced machine learning techniques for object detection, allows this work. Furthermore, using a multi-agent architecture allows the system to be modular, with the possibility of the different parts of the system working together, balancing the workload. The proposed system has been validated by testing it in different towns in Spain. The system has shown promising results in performing this task, with an F1-Score of 97.1%. [ABSTRACT FROM AUTHOR]

The electric energy production would be much more efficient if accurate estimations of the future demand were available, since these would allow allocating only the resources needed for the production of the right amount of energy required. With this motivation in mind, we propose a strategy, based on neuroevolution, that can be used to this aim. Our proposal uses a genetic algorithm in order to find a sub-optimal set of hyper-parameters for configuring a deep neural network, which can then be used for obtaining the forecasting. Such a strategy is justified by the observation that the performances achieved by deep neural networks are strongly dependent on the right setting of the hyper-parameters, and genetic algorithms have shown excellent search capabilities in huge search spaces. Moreover, we base our proposal on a distributed computing platform, which allows its use on a large time-series. In order to assess the performances of our approach, we have applied it to a large dataset, related to the electric energy consumption registered in Spain over almost 10 years. Experimental results confirm the validity of our proposal since it outperforms all other forecasting techniques to which it has been compared. [ABSTRACT FROM AUTHOR]

The automation of honey pollen visual sorting overcomes the limitations of the conventional procedure helping the specialist in this time-consuming task. In this work, a novel and comprehensive Ground Truth of almost 19,000 images (from optical microscopy) of the 16 most abundant types of grains/pollen particles present in citrus and rosemary honey from Spain was constructed. This task was assisted by a HoneyApp (also developed herein) for the labelling and annotation process. Subsequently, the effectiveness of different pre-existing automatic pollen recognizers based on convolutional neural networks (CNN) (VGG16, VGG19, InceptionV3, Xception, ResNet50, DenseNet201, MobileNetV2 and EfficientNetV2M) was tested together with a new network proposed in this paper (PolleNetV1). The extreme complexity of those pre-existing CNN and extensive use of millions of parameters makes this new proposal especially promising. Although with a slightly lower accuracy (average 96%) in determining the relative frequencies of different types of pollen grains/particles, it has considerable advantages such as simplicity and ability to be included in the future functionality to automate pollen recognition in honey. This is the first step to finally achieving an objective tool that allows the correct labelling of any types of pollen in honey, thus contributing to its transparency in the market. • A pollen images Ground Truth dataset from citrus and rosemary honeys was built. • A specific application (HoneyApp) to label/annotate pollen images was developed. • CNN networks were effective in counting the frequencies of pollen types in hone. • A new simpler, low memory and robust network (PolleNetV1) was created. • PolleNetV1 is expected to produce promising outcomes for other monoflorals. [ABSTRACT FROM AUTHOR]

In this paper we propose and discuss different Deep Learning-based ensemble algorithms for a problem of low-visibility events prediction due to fog. Specifically, seven different Deep Learning (DL) architectures have been considered, from which multiple individual learners are generated. Hyperparameters of the models, including parameters concerning data preprocessing, models architecture and training procedure, are randomly selected for each model within a pre-defined discrete range. Also, every model is trained with slightly different data sampled randomly, assuring that every models introduce variety in the ensemble. Then, three different information fusion techniques are employed to build the ensemble models. The influence of the filtering process and the elitism level (the percentage of the individual models entering the ensemble) is also assessed. The performance of the proposed methodology have been tested in two real problems of low-visibility events prediction due to orographical and radiation fog, at the north of Spain. Comparison with different Machine Learning, alternative DL algorithms and meteorological-based methods show the good performance of the proposed deep learning ensembles in this problem. [ABSTRACT FROM AUTHOR]

An ongoing outbreak of monkeypox virus (MPXV) was first reported in the United Kingdom on 6 May 2022. As of 17 November, there had been a total of 80 221 confirmed MPXV cases in over 110 countries. Based on data reported between 6 May and 30 June 2022 in the United Kingdom, Spain, and Germany, we applied a deep learning approach using convolutional neural networks to evaluate the parameters of the 2022 MPXV outbreak. The basic reproduction number (R0) of MPXV was estimated to be 2.32 in the United Kingdom, which indicates the active diffusion of MPXV since the beginning of the outbreak. The data from Spain and Germany produced higher median R0 values of 2.42 and 2.88, respectively. Importantly, the estimated R0 of MPXV in the three countries tends to the previously calculated R0 of smallpox (3.50 to 6.00). Furthermore, the incubation (1/ε) and infectious (1/γ) period was predicted between 9 and 10 days and 4−5 days, respectively. The R0 value derived from MPXV is consistent with the significantly increasing number of cases, indicating the risk of a rapid spread of MPXV worldwide, which would provide important insights for the prevention and control of MPXV epidemic. [ABSTRACT FROM AUTHOR]

In this paper, we present a new automatic screening method to assess whether a patient from ambulatory care or emergency should be referred to a cardiology service. This method is based on deep neural networks with pretraining and takes as an input a raw ECG signal without annotation. This work is based on a prospective clinical study that took place at Hospital Clínic in Barcelona between 2011–2012 and recruited 1390 patients. For each patient, we recorded a 12-lead ECG and the diagnosis was conducted by the cardiology service at the same hospital. Normal, borderline normal and normal variant ECGs were labelled as normal and the rest as abnormal . Our deep neural networks with pretraining were tested through cross-validation with a cohort of 416 test patients. The performance of our model was compared against other standard classification methods such as neural networks without pretraining, Support Vector Machines, Extreme Learning Machines, k-Nearest Neighbours and a professional classification algorithm certified for medical use that annotates the raw ECG signals prior to classification. The resulting best classifier was a pretrained neural network with three hidden layers and 700 units in every layer. This network yielded an accuracy of 0.8552, a sensitivity of 0.9176 and a specificity of 0.7827. The best alternative classification method was a Support Vector Machine with a Gaussian kernel, which yielded an accuracy of 0.8476, a sensitivity of 0.9446 and a specificity of 0.7346. The professional classification algorithm yielded an accuracy of 0.8407, a sensitivity of 0.8558 and a specificity of 0.8214. Neural networks with pretraining automatically obtain a representation of the input data without resorting to any annotation and, thus, simplify the process of assessing normality of ECG signals. The results that we have obtained are slightly better than those obtained with the professional classification system and, for some network configurations, they can be considered as exchangeable. Neural networks with pretraining open up a promising line of research for the automatic assessment of ECG signals that may be used in the future in clinical practice. [ABSTRACT FROM AUTHOR]

Most artificial intelligence (AI) research and innovations have concentrated in high-income countries, where imaging data, IT infrastructures and clinical expertise are plentiful. However, slower progress has been made in limited-resource environments where medical imaging is needed. For example, in Sub-Saharan Africa, the rate of perinatal mortality is very high due to limited access to antenatal screening. In these countries, AI models could be implemented to help clinicians acquire fetal ultrasound planes for the diagnosis of fetal abnormalities. So far, deep learning models have been proposed to identify standard fetal planes, but there is no evidence of their ability to generalise in centres with low resources, i.e. with limited access to high-end ultrasound equipment and ultrasound data. This work investigates for the first time different strategies to reduce the domain-shift effect arising from a fetal plane classification model trained on one clinical centre with high-resource settings and transferred to a new centre with low-resource settings. To that end, a classifier trained with 1792 patients from Spain is first evaluated on a new centre in Denmark in optimal conditions with 1008 patients and is later optimised to reach the same performance in five African centres (Egypt, Algeria, Uganda, Ghana and Malawi) with 25 patients each. The results show that a transfer learning approach for domain adaptation can be a solution to integrate small-size African samples with existing large-scale databases in developed countries. In particular, the model can be re-aligned and optimised to boost the performance on African populations by increasing the recall to 0.92 ± 0.04 and at the same time maintaining a high precision across centres. This framework shows promise for building new AI models generalisable across clinical centres with limited data acquired in challenging and heterogeneous conditions and calls for further research to develop new solutions for the usability of AI in countries with fewer resources and, consequently, in higher need of clinical support. [ABSTRACT FROM AUTHOR]

Accurate and rapid forecasting of short-term loads facilitates demand-side management by electricity retailers. The complexity of customer demand makes traditional forecasting methods incapable of meeting the accuracy requirements, so a self-attention based short-term load forecasting (STLF) considering demand-side management is proposed. In the data preprocessing stage, non-parametric kernel density estimation is used to construct customer electricity consumption feature curves, and then historical load data are used to delineate the feasible domain range for outlier detection. In the feature selection stage, the feature data are selected using variational modal decomposition and a maximum information coefficient to enhance the model prediction accuracy. In the model prediction stage, the decomposed intrinsic mode function components are independently predicted and reconstructed using an Informer based on improved self-attention. Additionally, the novel AdaBlief optimizer is used to optimize the model parameters. Cross-sectional and longitudinal experiments are conducted on a regional-level load dataset set in Spain. The experimental results prove that the proposed method is superior to other methods in STLF. [ABSTRACT FROM AUTHOR]

Time series forecasting of passenger demand is crucial for optimal planning of limited resources. For smart cities, passenger transport in urban areas is an increasingly important problem, because the construction of infrastructure is not the solution and the use of public transport should be encouraged. One of the most sophisticated techniques for time series forecasting is Long Short Term Memory (LSTM) neural networks. These deep learning models are very powerful for time series forecasting but are not interpretable by humans (black-box models). Our goal was to develop a predictive and linguistically interpretable model, useful for decision making using large volumes of data from different sources. Our case study was one of the most demanded bus lines of Madrid. We obtained an interpretable model from the LSTM neural network using a surrogate model and the 2-tuple fuzzy linguistic model, which improves the linguistic interpretability of the generated Explainable Artificial Intelligent (XAI) model without losing precision. [ABSTRACT FROM AUTHOR]

COVID-19 has caused millions of infections and deaths over the last 2 years. Machine learning models have been proposed as an alternative to conventional epidemiologic models in an effort to optimize short- and medium-term forecasts that will help health authorities to optimize the use of policies and resources to tackle the spread of the SARS-CoV-2 virus. Although previous machine learning models based on time pattern analysis for COVID-19 sensed data have shown promising results, the spread of the virus has both spatial and temporal components. This manuscript proposes a new deep learning model that combines a time pattern extraction based on the use of a Long-Short Term Memory (LSTM) Recurrent Neural Network (RNN) over a preceding spatial analysis based on a Convolutional Neural Network (CNN) applied to a sequence of COVID-19 incidence images. The model has been validated with data from the 286 health primary care centers in the Comunidad de Madrid (Madrid region, Spain). The results show improved scores in terms of both root mean square error (RMSE) and explained variance (EV) when compared with previous models that have mainly focused on the temporal patterns and dependencies. [ABSTRACT FROM AUTHOR]

Copyright of EMPIRIA: Revista de Metodología de Ciencias Sociales is the property of Editorial UNED and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

The increasing phenomenon of "cyberhate" is concerning because of the potential social implications of this form of verbal violence, which is aimed at already-stigmatized social groups. According to information collected by the Ministry of the Interior of Spain, the category of sexual orientation and gender identity is subject to the third-highest number of registered hate crimes, ranking behind racism/xenophobia and ideology. However, most of the existing computational approaches to online hate detection simultaneously attempt to address all types of discrimination, leading to weaker prototype performances. These approaches focus on other reasons for hate--primarily racism and xenophobia--and usually focus on English messages. Furthermore, few detection models have used manually generated databases as a training corpus. Using supervised machine learning techniques, the present research sought to overcome these limitations by developing and evaluating an automatic detector of hate speech motivated by gender and sexual orientation. The focus was Spanish-language posts on Twitter. For this purpose, eight predictive models were developed from an ad hoc generated training corpus, using shallow modeling and deep learning. The evaluation metrics showed that the deep learning algorithm performed significantly better than the shallow modeling algorithms, and logistic regression yielded the best performance of the shallow algorithms. [ABSTRACT FROM AUTHOR]

Forecasting the evolution of contagion dynamics is still an open problem to which mechanistic models only offer a partial answer. To remain mathematically or computationally tractable, these models must rely on simplifying assumptions, thereby limiting the quantitative accuracy of their predictions and the complexity of the dynamics they can model. Here, we propose a complementary approach based on deep learning where effective local mechanisms governing a dynamic on a network are learned from time series data. Our graph neural network architecture makes very few assumptions about the dynamics, and we demonstrate its accuracy using different contagion dynamics of increasing complexity. By allowing simulations on arbitrary network structures, our approach makes it possible to explore the properties of the learned dynamics beyond the training data. Finally, we illustrate the applicability of our approach using real data of the COVID-19 outbreak in Spain. Our results demonstrate how deep learning offers a new and complementary perspective to build effective models of contagion dynamics on networks. Prediction of contagion dynamics is of relevance for epidemic and social complex networks. Murphy et al. propose a data-driven approach based on deep learning which allows to learn mechanisms governing network dynamics and make predictions beyond the training data for arbitrary network structures. [ABSTRACT FROM AUTHOR]

Time series forecasting of disaggregated freight flow is a key issue in decision-making by port authorities. For this purpose and to test new deep learning techniques we have selected seven time series of imported goods from Morocco to Spain through the port of Algeciras, and we have tested two forecasting deep neural networks models: dilated causal convolutional and encoder–decoder recurrent. We have experimented with four different granularities for each series: quarterly, monthly, weekly and daily. The results show that our neural network models can manage these raw series without first removing seasonality or trend. We also highlight the ability of neural models to work with a fixed input size of one year, being able to make good predictions using the same input size for all granularities. The two deep learning models have globally improved the benchmarks of the M4 Competition of forecasting. Each neural network model obtains its best results under different circumstances: the recurrent one with daily granularity and intermittent series, and the convolutional one with weekly and monthly granularities. [ABSTRACT FROM AUTHOR]

The Covid-19 epidemic poses a serious public health threat to the world, where people with little or no pre-existing human immunity can be more vulnerable to its effects. Thus, developing surveillance systems for predicting the Covid-19 pandemic at an early stage could save millions of lives. In this study, a deep learning algorithm and a Holt-trend model are proposed to predict the coronavirus. The Long-Short Term Memory (LSTM) and Holt-trend algorithms were applied to predict conirmed numbers and death cases. The real time data used has been collected from the World Health Organization (WHO). In the proposed research, we have considered three countries to test the proposed model, namely Saudi Arabia, Spain and Italy. The results suggest that the LSTM models show better performance in predicting the cases of coronavirus patients. Standard measure performance Mean squared Error (MSE), Root Mean Squared Error (RMSE), Mean error and correlation are employed to estimate the results of the proposed models. The empirical results of the LSTM, using the correlation metrics, are 99.94%, 99.94% and 99.91% in predicting the number of confirmed cases in the three countries. As far as the results of the LSTM model in predicting the number of death of Covid-19, they are 99.86%, 98.876% and 99.16% with respect to Saudi Arabia, Italy and Spain respectively. Similarly, the experiment's results of the Holt-Trend model in predicting the number of confirmed cases of Covid-19, using the correlation metrics, are 99.06%, 99.96%and 99.94%,whereas the results of the Holt-Trend model in predicting the number of death cases are 99.80%, 99.96%and 99.94% with respect to the Saudi Arabia, Italy and Spain respectively. The empirical results indicate the efficient performance of the presented model in predicting the number of confirmed and death cases of Covid-19 in these countries. Such findings provide better insights regarding the future of Covid-19 this pandemic in general. The results were obtained by applying time series models, which need to be considered for the sake of saving the lives of many people. [ABSTRACT FROM AUTHOR]

Featured Application: Energy demand forecasting to improve power generation management. Modern energy systems collect high volumes of data that can provide valuable information about energy consumption. Electric companies can now use historical data to make informed decisions on energy production by forecasting the expected demand. Many deep learning models have been proposed to deal with these types of time series forecasting problems. Deep neural networks, such as recurrent or convolutional, can automatically capture complex patterns in time series data and provide accurate predictions. In particular, Temporal Convolutional Networks (TCN) are a specialised architecture that has advantages over recurrent networks for forecasting tasks. TCNs are able to extract long-term patterns using dilated causal convolutions and residual blocks, and can also be more efficient in terms of computation time. In this work, we propose a TCN-based deep learning model to improve the predictive performance in energy demand forecasting. Two energy-related time series with data from Spain have been studied: the national electric demand and the power demand at charging stations for electric vehicles. An extensive experimental study has been conducted, involving more than 1900 models with different architectures and parametrisations. The TCN proposal outperforms the forecasting accuracy of Long Short-Term Memory (LSTM) recurrent networks, which are considered the state-of-the-art in the field. [ABSTRACT FROM AUTHOR]

Copyright of Revista Facultad de Ingeniería Universidad de Antioquia is the property of Universidad de Antioquia and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

The road surface area extraction task is generally carried out via semantic segmentation over remotely-sensed imagery. However, this supervised learning task is often costly as it requires remote sensing images labelled at the pixel level, and the results are not always satisfactory (presence of discontinuities, overlooked connection points, or isolated road segments). On the other hand, unsupervised learning does not require labelled data and can be employed for post-processing the geometries of geospatial objects extracted via semantic segmentation. In this work, we implement a conditional Generative Adversarial Network to reconstruct road geometries via deep inpainting procedures on a new dataset containing unlabelled road samples from challenging areas present in official cartographic support from Spain. The goal is to improve the initial road representations obtained with semantic segmentation models via generative learning. The performance of the model was evaluated on unseen data by conducting a metrical comparison where a maximum Intersection over Union (IoU) score improvement of 1.3% was observed when compared to the initial semantic segmentation result. Next, we evaluated the appropriateness of applying unsupervised generative learning using a qualitative perceptual validation to identify the strengths and weaknesses of the proposed method in very complex scenarios and gain a better intuition of the model's behaviour when performing large-scale post-processing with generative learning and deep inpainting procedures and observed important improvements in the generated data. [ABSTRACT FROM AUTHOR]

Modelling fire behaviour in forest fires is based on meteorological, topographical, and vegetation data, including species' type. To accurately parameterise these models, an inventory of the area of analysis with the maximum spatial and temporal resolution is required. This study investigated the use of UAV-based digital aerial photogrammetry (UAV-DAP) point clouds to classify tree and shrub species in Mediterranean forests, and this information is key for the correct generation of wildfire models. In July 2020, two test sites located in the Natural Park of Sierra Calderona (eastern Spain) were analysed, registering 1036 vegetation individuals as reference data, corresponding to 11 shrub and one tree species. Meanwhile, photogrammetric flights were carried out over the test sites, using a UAV DJI Inspire 2 equipped with a Micasense RedEdge multispectral camera. Geometrical, spectral, and neighbour-based features were obtained from the resulting point cloud generated. Using these features, points belonging to tree and shrub species were classified using several machine learning methods, i.e., Decision Trees, Extra Trees, Gradient Boosting, Random Forest, and MultiLayer Perceptron. The best results were obtained using Gradient Boosting, with a mean cross-validation accuracy of 81.7% and 91.5% for test sites 1 and 2, respectively. Once the best classifier was selected, classified points were clustered based on their geometry and tested with evaluation data, and overall accuracies of 81.9% and 96.4% were obtained for test sites 1 and 2, respectively. Results showed that the use of UAV-DAP allows the classification of Mediterranean tree and shrub species. This technique opens a wide range of possibilities, including the identification of species as a first step for further extraction of structure and fuel variables as input for wildfire behaviour models. [ABSTRACT FROM AUTHOR]

This work proposes a semi-parametric approach to estimate the evolution of COVID-19 (SARS-CoV-2) in Spain. Considering the sequences of 14-day cumulative incidence of all Spanish regions, it combines modern Deep Learning (DL) techniques for analyzing sequences with the usual Bayesian Poisson-Gamma model for counts. The DL model provides a suitable description of the observed time series of counts, but it cannot give a reliable uncertainty quantification. The role of expert elicitation of the expected number of counts and its reliability is DL predictions' role in the proposed modelling approach. Finally, the posterior predictive distribution of counts is obtained in a standard Bayesian analysis using the well known Poisson-Gamma model. The model allows to predict the future evolution of the sequences on all regions or estimates the consequences of eventual scenarios. [ABSTRACT FROM AUTHOR]

This work proposes a quantile regression neural network based on a novel constrained weighted quantile loss (CWQLoss) and its application to probabilistic short and medium-term electric-load forecasting of special interest for smart grids operations. The method allows any point forecast neural network based on a multivariate multi-output regression model to be expanded to become a quantile regression model. CWQLoss extends the pinball loss to more than one quantile by creating a weighted average for all predictions in the forecast window and across all quantiles. The pinball loss for each quantile is evaluated separately. The proposed method imposes additional constraints on the quantile values and their associated weights. It is shown that these restrictions are important to have a stable and efficient model. Quantile weights are learned end-to-end by gradient descent along with the network weights. The proposed model achieves two objectives: (a) produce probabilistic (quantile and interval) forecasts with an associated probability for the predicted target values. (b) generate point forecasts by adopting the forecast for the median (0.5 quantiles). We provide specific metrics for point and probabilistic forecasts to evaluate the results considering both objectives. A comprehensive comparison is performed between a selection of classic and advanced forecasting models with the proposed quantile forecasting model. We consider different scenarios for the duration of the forecast window (1 h, 1-day, 1-week, and 1-month), with the proposed model achieving the best results in almost all scenarios. Additionally, we show that the proposed method obtains the best results when an additive ensemble neural network is used as the base model. The experimental results are drawn from real loads of a medium-sized city in Spain. [ABSTRACT FROM AUTHOR]

This study aims to produce accurate predictions of the NO2 concentrations at a specific station of a monitoring network located in the Bay of Algeciras (Spain). Artificial neural networks (ANNs) and sequence-to-sequence long short-term memory networks (LSTMs) were used to create the forecasting models. Additionally, a new prediction method was proposed combining LSTMs using a rolling window scheme with a cross-validation procedure for time series (LSTM-CVT). Two different strategies were followed regarding the input variables: using NO2 from the station or employing NO2 and other pollutants data from any station of the network plus meteorological variables. The ANN and LSTM-CVT exogenous models used lagged datasets of different window sizes. Several feature ranking methods were used to select the top lagged variables and include them in the final exogenous datasets. Prediction horizons of t + 1, t + 4 and t + 8 were employed. The exogenous variables inclusion enhanced the model's performance, especially for t + 4 (ρ ≈ 0.68 to ρ ≈ 0.74) and t + 8 (ρ ≈ 0.59 to ρ ≈ 0.66). The proposed LSTM-CVT method delivered promising results as the best performing models per prediction horizon employed this new methodology. Additionally, per each parameter combination, it obtained lower error values than ANNs in 85% of the cases. [ABSTRACT FROM AUTHOR]

Improving road safety through artificial intelligence-based systems is now crucial turning smart cities into a reality. Under this highly relevant and extensive heading, an approach is proposed to improve vehicle detection in smart crosswalks using machine learning models. Contrarily to classic fuzzy classifiers, machine learning models do not require the readjustment of labels that depend on the location of the system and the road conditions. Several machine learning models were trained and tested using real traffic data taken from urban scenarios in both Portugal and Spain. These include random forest, time-series forecasting, multi-layer perceptron, support vector machine, and logistic regression models. A deep reinforcement learning agent, based on a state-of-the-art double-deep recurrent Q-network, is also designed and compared with the machine learning models just mentioned. Results show that the machine learning models can efficiently replace the classic fuzzy classifier. [ABSTRACT FROM AUTHOR]