Search

• Processing multiple EEGs simultaneously. • Automatic feature extraction without further filtering. • Deeply comparison of the number of trainable parameters. Nowadays, among the Deep Learning works, there is a tendency to develop networks with millions of trainable parameters. However, this tendency has two main drawbacks: overfitting and resource consumption due to the low-quality features extracted by those networks. This paper presents a study focused on the scoring of sleeping EEG signals to measure if the increase of the pressure on the features due to a reduction of the number though different techniques results in a benefit. The work also studies the convenience of increasing the number of input signals in order to allow the network to extract better features. Additionally, it might be highlighted that the presented model achieves comparable results to the state-of-the-art with 1000 times less trainable and the presented model uses the whole dataset instead of the simplified versions in the published literature. [ABSTRACT FROM AUTHOR]

Genetic Programming (GP) is a technique which is able to solve different problems through the evolution of mathematical expressions. However, in order to be applied, its tendency to overfit the data is one of its main issues. The use of a validation dataset is a common alternative to prevent overfitting in many Machine Learning (ML) techniques, including GP. But, there is one key point which differentiates GP and other ML techniques: instead of training a single model, GP evolves a population of models. Therefore, the use of the validation dataset has several possibilities because any of those evolved models could be evaluated. This work explores the possibility of using the validation dataset not only on the training-best individual but also in a subset with the training-best individuals of the population. The study has been conducted with 5 well-known databases performing regression or classification tasks. In most of the cases, the results of the study point out to an improvement when the validation dataset is used on a subset of the population instead of only on the training-best individual, which also induces a reduction on the number of nodes and, consequently, a lower complexity on the expressions. [ABSTRACT FROM AUTHOR]

Sleeping problems have become one of the major diseases all over the world. To tackle this issue, the basic tool used by specialists is the polysomnogram, which is a collection of different signals recorded during sleep. After its recording, the specialists have to score the different signals according to one of the standard guidelines. This process is carried out manually, which can be a high-time-consuming task and very prone to annotation errors. Therefore, over the years, many approaches have been explored in an attempt to support the specialists in this task. In this paper, an approach based on convolutional neural networks is presented, where an in-depth comparison is made in order to determine the convenience of using more than one signal simultaneously as input. This approach is similar to the one made in other problems although, additionally to those models, they were also used as parts of an ensemble model to check whether any useful information can be extracted from processing a single signal at a time which the dual-signal model cannot identify. Tests have been performed by using a well-known dataset called sleep-EDF-expanded, which is the most commonly used dataset as benchmark for this problem. The tests were carried out with a leave-one-out cross-validation over the patients, which ensures that there is no possible contamination between training and testing. The resulting proposal is a network smaller than previously published ones, but it overcomes the results of any previous models on the same dataset. The best result shows an accuracy of 92.67% and a Cohen's kappa value over 0.84 compared to human experts. [ABSTRACT FROM AUTHOR]

Facial biometrics play an essential role in the fields of law enforcement and forensic sciences. When comparing facial traits for human identification in photographs or videos, the analysis must account for several factors that impair the application of common identification techniques, such as illumination, pose, or expression. In particular, facial attributes can drastically change depending on the distance between the subject and the camera at the time of the picture. This effect is known as perspective distortion, which can severely affect the outcome of the comparative analysis. Hence, knowing the subject-to-camera distance of the original scene where the photograph was taken can help determine the degree of distortion, improve the accuracy of computer-aided recognition tools, and increase the reliability of human identification and further analyses. In this paper, we propose a deep learning approach to estimate the subject-to-camera distance of facial photographs: FacialSCDnet. Furthermore, we introduce a novel evaluation metric designed to guide the learning process, based on changes in facial distortion at different distances. To validate our proposal, we collected a novel dataset of facial photographs taken at several distances using both synthetic and real data. Our approach is fully automatic and can provide a numerical distance estimation for up to six meters, beyond which changes in facial distortion are not significant. The proposed method achieves an accurate estimation, with an average error below 6 cm of subject-to-camera distance for facial photographs in any frontal or lateral head pose, robust to facial hair, glasses, and partial occlusion. • Accurate estimation of subject-to-camera distance in portrait photographs. • A novel metric is proposed, based on the effects of perspective in facial distortion. • A new database of facial images at a distance is introduced for human identification. • A transfer learning approach overcomes the limitations of current methods. • Robust to expression, occlusion and pose without requiring anatomical information. [ABSTRACT FROM AUTHOR]

Based on a solid mathematical background, this paper proposes a method for Symbolic Regression that enables the extraction of mathematical expressions from a dataset. Contrary to other approaches, such as Genetic Programming, the proposed method is deterministic and, consequently, does not require the creation of a population of initial solutions. Instead, a simple expression is grown until it fits the data. This method has been compared with four well-known Symbolic Regression techniques with a large number of datasets. As a result, on average, the proposed method returns better performance than the other techniques, with the advantage of returning mathematical expressions that can be easily used by different systems. Additionally, this method makes it possible to establish a threshold at the complexity of the expressions generated, i.e., the system can return mathematical expressions that are easily analyzed by the user, as opposed to other techniques that return very large expressions. • Novel technique for Machine Learning and Symbolic Regression. • Possibility to explicitly set the complexity of the developed expression. • Results compared to other Symbolic Regression techniques and better on a large number of datasets. • Much simpler expressions than in GP. [ABSTRACT FROM AUTHOR]

• ML Models for HAB prediction and shellfish farming management decision support. • Study of the importance of variables involved in the forecasting of HAB. • Analysing the behaviour of machine learning models against unbalanced datasets. • Prediction of production cessation by HAB with 91.83% accuracy and 97.34% recall. Mussel farming is one of the most important aquaculture industries. The main risk to mussel farming is harmful algal blooms (HABs), which pose a risk to human consumption. In Galicia, the Spanish main producer of cultivated mussels, the opening and closing of the production areas is controlled by a monitoring program. In addition to the closures resulting from the presence of toxicity exceeding the legal threshold, in the absence of a confirmatory sampling and the existence of risk factors, precautionary closures may be applied. These decisions are made by experts without the support or formalisation of the experience on which they are based. Therefore, this work proposes a predictive model capable of supporting the application of precautionary closures. Achieving sensitivity, accuracy and kappa index values of 97.34%, 91.83% and 0.75 respectively, the kNN algorithm has provided the best results. This allows the creation of a system capable of helping in complex situations where forecast errors are more common. [ABSTRACT FROM AUTHOR]

Abstract: The recognition of seizures is very important for the diagnosis of patients with epilepsy. The seizure is a process of rhythmic discharge in brain and occurs rarely and unpredictably. This behavior generates a need of an automatic detection of seizures by using the signals of long-term electroencephalographic (EEG) recordings. Due to the non-stationary character of EEG signals, the conventional methods of frequency analysis are not the best alternative to obtain good results in diagnostic purpose. The present work proposes a method of EEG signal analysis based on star graph topological indices (SGTIs) for the first time. The signal information, such as amplitude and time occurrence, is codified into invariant SGTIs which are the basis for the classification models that can discriminate the epileptic EEG records from the non-epileptic ones. The method with SGTIs and the simplest linear discriminant methods provide similar results to those previously published, which are based on the time-frequency analysis and artificial neural networks. Thus, this work proposes a simpler and faster alternative for automatic detection of seizures from the EEG recordings. [Copyright &y& Elsevier]

Currently, consumers are more concerned about freshness and quality of food. Poultry egg storage time is a freshness and quality indicator in industrial and consumer applications, even though egg marking is not always required outside the European Union. Other authors have already published works using expensive laboratory equipment in order to determine the storage time and freshness of eggs. This paper presents a novel alternative method based on low-cost devices for the rapid and non-destructive prediction of egg storage time at room temperature (23 ± 1 °C). H&N brown flock with 49-week-old hens were used as a source for the sampled eggs. Samples were scanned for a period of 22 days beginning from the time the egg was laid. The spectral acquisition was performed using a low-cost near-infrared reflectance (NIR) spectrometer which has a wavelength range between 740 nm and 1070 nm. The resulting dataset of 660 samples was randomly split according to a 10-fold cross-validation in order to be used in a contrast and optimization process of two machine learning algorithms. During the optimization, several models were tested to develop a robust calibration model. The best model used a Savitzky Golay pre-processing technique with a third derivative order and an artificial neural network with ten neurons in one hidden layer. Regressing the storage time of the eggs, tests achieved a coefficient of determination ( R-squared ) of 0.8319 ± 0.0377 and a root mean squared error in cross-validation test set ( RMSECV ) of 1.97 days. Although further work is needed, this technique shows industrial potential and consumer utility to determine an egg's freshness using a low-cost spectrometer connected to a smartphone. [ABSTRACT FROM AUTHOR]

In recent years, mainly due to the application of smartphones in this area, research in human activity recognition (HAR) has shown a continuous and steady growth. Thanks to its wide range of sensors, its size, its ease of use, its low price and its applicability in many other fields, it is a highly attractive option for researchers. However, the vast majority of studies carried out so far focus on laboratory settings, outside of a real-life environment. In this work, unlike in other papers, progress was sought on the latter point. To do so, a dataset already published for this purpose was used. This dataset was collected using the sensors of the smartphones of different individuals in their daily life, with almost total freedom. To exploit these data, numerous experiments were carried out with various machine learning techniques and each of them with different hyperparameters. These experiments proved that, in this case, tree-based models, such as Random Forest, outperform the rest. The final result shows an enormous improvement in the accuracy of the best model found to date for this purpose, from 74.39% to 92.97%. • This paper presents a comparison of the main machine learning algorithms for HAR. • A dataset taken in a real-life environment was used, unlike in other studies. • Experiments were done to get the best model configurations for long-themed activities. • This work also reviews the gyroscope's real influence to the final results. • The current approaches in this field, oriented towards real life, were highly improved. [ABSTRACT FROM AUTHOR]

Path Planning problems with Unmanned Aerial Vehicles (UAVs) are among the most studied knowledge areas in the related literature. However, few of them have been applied to groups of UAVs. The use of swarms allows to speed up the flight time and, thus, reducing the operational costs. When combined with Artificial Intelligence (AI) algorithms, a single system or operator can control all aircraft while optimal paths for each one can be computed. In order to introduce the current situation of these AI-based systems, a review of the most novel and relevant articles was carried out. This review was performed in two steps: first, a summary of the found articles; second, a quantitative analysis of the publications found based on different factors, such as the temporal evolution or the number of articles found based on different criteria. Therefore, this review provides not only a summary of the most recent work but it gives an overview of the trend in the use of AI algorithms in UAV swarms for Path Planning problems. The AI techniques of the articles found can be separated into four main groups based on their technique: reinforcement Learning techniques, Evolutive Computing techniques, Swarm Intelligence techniques, and, Graph Neural Networks. The final results show an increase in publications in recent years and that there is a change in the predominance of the most widely used techniques. [ABSTRACT FROM AUTHOR]

Path Planning methods for the autonomous control of Unmanned Aerial Vehicle (UAV) swarms are on the rise due to the numerous advantages they bring. There are increasingly more scenarios where autonomous control of multiple UAVs is required. Most of these scenarios involve a large number of obstacles, such as power lines or trees. Despite these challenges, there are also several advantages; if all UAVs can operate autonomously, personnel expenses can be reduced. Additionally, if their flight paths are optimized, energy consumption is reduced, leaving more battery time for other operations. In this paper, a Reinforcement Learning-based system is proposed to solve this problem in environments with obstacles by utilizing Q-Learning. This method allows a model, in this case, an Artificial Neural Network, to self-adjust by learning from its mistakes and successes. Regardless of the map's size or the number of UAVs in the swarm, the goal of these paths is to ensure complete coverage of an area with fixed obstacles for tasks like field prospecting. Setting goals or having any prior information apart from the provided map is not required. During the experimentation phase, five maps of varying sizes were used, each with different obstacles and a varying number of UAVs. To evaluate the quality of the results, the number of actions taken by each UAV to complete the task in each experiment was considered. The results indicate that the system achieves solutions with fewer movements as the number of UAVs increases. An increasing number of UAVs on a map lead to solutions in fewer moves. The results have been compared, and a statistical significance analysis has been conducted on the proposed model's outcomes, demonstrating its capabilities. Thus, it is shown that a two-layer Artificial Neural Network used to implement a Q-Learning algorithm is sufficient to operate on maps with obstacles. • Q-Learning for UAV swarm to determine the best flight path to cover the most area. • Estimate the flight paths on any size map with obstacles and without map information. • A system capable of calculating paths without the need for a smoothing stage. • Statistical comparison of a single ANN for each UAV against a global ANN for all UAVs. • An aircraI architecture-independent path optimization criterion. [ABSTRACT FROM AUTHOR]

In recent years, human activity recognition has become a hot topic inside the scientific community. The reason to be under the spotlight is its direct application in multiple domains, like healthcare or fitness. Additionally, the current worldwide use of smartphones makes it particularly easy to get this kind of data from people in a non-intrusive and cheaper way, without the need for other wearables. In this paper, we introduce our orientation-independent, placement-independent and subject-independent human activity recognition dataset. The information in this dataset is the measurements from the accelerometer, gyroscope, magnetometer, and GPS of the smartphone. Additionally, each measure is associated with one of the four possible registered activities: inactive, active, walking and driving. This work also proposes asupport vector machine (SVM) model to perform some preliminary experiments on the dataset. Considering that this dataset was taken from smartphones in their actual use, unlike other datasets, the development of a good model on such data is an open problem and a challenge for researchers. By doing so, we would be able to close the gap between the model and a real-life application. [ABSTRACT FROM AUTHOR]

Harmful algal blooms (HABs) are episodes of high concentrations of algae that are potentially toxic for human consumption. Mollusc farming can be affected by HABs because, as filter feeders, they can accumulate high concentrations of marine biotoxins in their tissues. To avoid the risk to human consumption, harvesting is prohibited when toxicity is detected. At present, the closure of production areas is based on expert knowledge and the existence of a predictive model would help when conditions are complex and sampling is not possible. Although the concentration of toxin in meat is the method most commonly used by experts in the control of shellfish production areas, it is rarely used as a target by automatic prediction models. This is largely due to the irregularity of the data due to the established sampling programs. As an alternative, the activity status of production areas has been proposed as a target variable based on whether mollusc meat has a toxicity level below or above the legal limit. This new option is the most similar to the actual functioning of the control of shellfish production areas. For this purpose, we have made a comparison between hybrid machine learning models like Neural-Network-Adding Bootstrap (BAGNET) and Discriminative Nearest Neighbor Classification (SVM-KNN) when estimating the state of production areas. The study has been carried out in several estuaries with different levels of complexity in the episodes of algal blooms to demonstrate the generalization capacity of the models in bloom detection. As a result, we could observe that, with an average recall value of 93.41% and without dropping below 90% in any of the estuaries, BAGNET outperforms the other models both in terms of results and robustness. • Presence of toxicity in molluscan shellfish meat as a target for predicting HABs. • Hybrid AI methods outperform classical methods in predicting HABs. • Analyzing the behavior of hybrid machine learning models against unbalanced datasets. • BAGNET offers the best results in the prediction of HABs in complex regions. [ABSTRACT FROM AUTHOR]