Search

Knihovna PaGMO nabízí několik optimizačních algoritmů pro stanovení optimálních parametrů black-boxmodelu. Takovým modelem může být například model pro homeostázu glukózy. Protože nás zajímá výpočet a predikce hladin glukózy u diabetických pacientů, testujeme knihovnu PaGMOpro tento konkrétní úkol. S využitím tří scénářů testujeme jednotlivé algoritmy PaGMO a porovnáváme je s našimi dřívějšími výsledky, které jsme získali pomocí de-randomizované meta diferenciální evoluce. Všechny testovací scénáře se zabývají skutečným případem zpracování signálu systému kontinuálního monitorování glukózy. Konkrétně se zabýváme rekonstrukcí a predikcí signálu. The PaGMO framework offers several optimizationalgorithms to determine optimal parameters of a black-boxmodel. Such a model could be, for example, that for glucose homeostasis. As we are concerned about calculating and predicting glucose levels for diabetic patients, we evaluate the PaGMO framework for this particular task. Using three scenarios, we test PaGMO’s individual algorithms and compare them to ourprevious results, which we obtained with de-randomized Meta-Differential Evolutions. All testing scenarios address real aspectsof processing a signal of the continuous glucose monitoring system. Specifically, we address signal reconstruction and prediction.

Tento článek ilustruje vývoj a použitelnost Evolučního Počítání pro zlepšení léčby pacientů s diabetem typu 1, kteří léčbu inzulínem. Jelikož je takové onemocnění spojeno se špatnou funkcí slinivky břišní, která produkuje nedostatečné množství inzulínu, způsob, jak zvýšit kvalitu života těchto pacientů, je vytvořit umělou slinivku schopnou uměle regulovat dávkování inzulínu. Cílem této práce je extrapolace regresního modelu, schopného odhadnout krevní glukózu (BG) pomocí měření intersticiální glukózy (IG) a jejich prvních derivací. V této studii demonstruujeme použitelnost tohoto přístupu. Zejména s ohledem na vysokou složitost vzájemných interakcí je navržena evoluční strategie, která extrapoluje matematický vztah mezi BG a IG a jeho derivací. Vypočítané modely jsou porovnány s jinými modely během experimentů na personalizovaných a globálních datech. This paper illustrates the development and the applicability of an Evolutionary Computation approach to enhance the treatment of Type-1 diabetic patients that necessitate insulin injections. In fact, being such a disease associated to a malfunctioning pancreas that generates an insufficient amount of insulin, a way to enhance the quality of life of these patients is to implement an artificial pancreas able to artificially regulate the insulin dosage. This work aims at extrapolating a regression model, capable of estimating the blood glucose (BG) through interstitial glucose (IG) measurements and their numerical first derivatives. Such an approach represents a viable preliminary stage in building the basic component of this artificial pancreas. In particular, considered the high complexity of the reciprocal interactions, an evolutionary-based strategy is outlined to extrapolate a mathematical relationship between BG and IG and its derivative. The investigation is carried out about the accuracy of personalized models and of a global relationship model for all of the subjects under examination. The discovered models are assessed through a comparison with other models during the experiments on personalized and global data.

Deep Neural Networks (DNNs) may be very effective for the classification over highly-sized data sets, especially in the medical domain, where the recognition of the occurrence of a specific event related to a disease is of high importance. Unfortunately, DNNs suffer from the drawback that a good set of values for their configuration hyper-parameters must be found. Currently, this is done through the use of either trial-and-error methods or sampling-based ones. In this paper we propose a new approach to find the most suitable structure for a DNN used for a classification problem in terms of achievement of the highest classification accuracy. This approach is based on a distributed version of Differential Evolution (DE), a variety of an Evolutionary Algorithm. To evaluate the approach, in this paper we investigate this issue with reference to Obstructive Sleep Apnea (OSA). OSA is an important medical problem consisting of episodes taking place during night in which a subject stops breathing due to a constriction of the upper airways. This deteriorates the quality of life and may have dangerous, and even lethal, consequences on both short and long term. An accurate classification is a very crucial step for the OSA treatment, because understanding automatically that a subject is experiencing such an episode may be decisive if prompt medical action is needed. In our experiments, classification takes place on a data set in which each item contains the values of 17 Heart Rate Variability parameters, extracted from ElectroCardiography signals, and the annotation of OSA events. We have extracted this data set from the real-world Sleep Heart Health Study database. The results obtained by the distributed DE are compared against those of the Grid Search as well as against those achieved by 13 well-known classification tools. The use of a distributed DE version turns out to be very effective in automatically obtaining DNN structures with higher classification accuracy with respect to Grid Search (72.95% versus 72.61%), and allows saving a high amount of time (three hours as opposed to 65 h and 40 min). Moreover, the proposed method outperforms in terms of higher accuracy all the other classifiers investigated, as it is evidenced also by statistical analysis. Numerically, the runner-up, i.e., JRip, achieves as its best value 72.01% and 71.50% on average over 25 runs, both values being lower than 72.95% and 72.74% obtained by our dDE.

A physiological model improves delivered healthcare, when constructing a medical device. Such a model comprises a number of parameters. While an analytical method determines model parameters, an evolutionary algorithm can improve them further. As evolutionary algorithms were designed on top of random-number generators, their results are not deterministic. This raises a concern about their applicability to medical devices. Medical-device algorithm must produce an output with a minimum guaranteed accuracy. Therefore, we applied de-randomized sequences to Meta-Differential Evolution instead of using a random-number generator. Eventually, we designed an optimization method based on zooming with derandomized sequences as an alternative to the Meta-Differential Evolution. As the experimental setup, we predicted glucose-level signal to cover a blind window of glucose-monitoring signal that results from a physiological lag in glucose transportation. Completely de-randomized differential evolution exhibited the same accuracy and precision as completely non-deterministic differential evolution. They produced 93% of glucose levels with relative error less than or equal to 15%. Fyziologický model zlepšuje poskytovanou zdravotní péči při konstrukci zdravotnického zařízení. Takový model zahrnuje řadu parametrů. Zatímco analytická metoda určuje parametry modelu, evoluční algoritmus je může dále vylepšovat. Protože vývojové algoritmy byly navrženy na generátorech náhodných čísel, jejich výsledky nejsou deterministické. To vzbuzuje obavy ohledně jejich použitelnosti na zdravotnické prostředky. Algoritmus zdravotnického zařízení musí produkovat výstup s minimální zaručenou přesností. Proto jsme použili de-randomizované sekvence na meta-diferenciální evoluci namísto použití generátoru náhodných čísel. Nakonec jsme navrhli optimalizační metodu založenou na zvětšování pomocí náhodných sekvencí jako alternativu k meta-diferenciální evoluci. Jako experimentální nastavení jsme předpovídali signál hladiny glukózy tak, aby pokrýval slepé okno signálu monitorování glukózy, které je výsledkem fyziologického zpoždění v transportu glukózy. Zcela de-randomizovaný diferenciální vývoj vykazoval stejnou přesnost a přesnost jako zcela nedeterministický diferenciální vývoj. Oba spočítali 93% koncentrací glukózy s relativní chybou menší nebo rovnou 15%.

Within this paper a general-purpose distributed evolutionary algorithm is presented, and is applied to the pair-wise registration of range images. Registration is carried out by utilizing the Grid Closest Point (GCP) for the graphical registration operations and the distributed algorithm to search for the best possible transformation of a scene image that, merged with the model image, yields a 3D reconstruction of the original object. The evolutionary algorithm is a distributed Differential Evolution algorithm that exploits an asynchronous migration mechanism and a multi-population recombination information exchange. Such an algorithm is provided with an adaptive updating scheme based on chaotic features for dynamically updating the control parameters. The scope of the paper is to speed up the registration process by using processor specialized to handle graphical operations and multi-core platforms. On the one hand, we investigate the use of either Graphic Processing Units (GPUs) or multi-core architectures to lower the execution time of the GCP procedure. On the other hand, we evaluate the performance of the distributed evolutionary algorithm in terms of solution quality by examining different multi-core architectures. Experimental results on a set of publicly available images show that, to perform the GCP, reductions in the execution times by one order of magnitude are obtained by harnessing the computational power of GPU and multi-core platforms with respect to the execution on a CPU-based framework. Furthermore, a comparison with the state-of-the-art sequential evolutionary algorithm for range image registration reveals that the adaptive distributed Differential Evolution algorithm allows attaining integral 3D models from 3D scan datasets that are better in terms of both quality and robustness.

In the last decades, early disease identification through non-invasive and automatic methodologies has gathered increasing interest from the scientific community. Among others, Parkinson's disease (PD) has received special attention in that it is a severe and progressive neuro-degenerative disease. As a consequence, early diagnosis would provide more effective and prompt care strategies, that cloud successfully influence patients’ life expectancy. However, the most performing systems implement the so called black-box approach, which do not provide explicit rules to reach a decision. This lack of interpretability, has hampered the acceptance of those systems by clinicians and their deployment on the field. In this context, we perform a thorough comparison of different machine learning (ML) techniques, whose classification results are characterized by different levels of interpretability. Such techniques were applied for automatically identify PD patients through the analysis of handwriting and drawing samples. Results analysis shows that white-box approaches, such as Cartesian Genetic Programming and Decision Tree, allow to reach a twofold goal: support the diagnosis of PD and obtain explicit classification models, on which only a subset of features (related to specific tasks) were identified and exploited for classification. Obtained classification models provide important insights for the design of non-invasive, inexpensive and easy to administer diagnostic protocols. Comparison of different ML approaches (in terms of both accuracy and interpretability) has been performed on the features extracted from the handwriting and drawing samples included in the publicly available PaHaW and NewHandPD datasets. The experimental findings show that the Cartesian Genetic Programming outperforms the white-box methods in accuracy and the black-box ones in interpretability.

Deep Neural Networks are being more and more widely used to perform several tasks over highly-sized datasets, one of them being classification. Finding good configurations for Deep Neural Network structures is a very important problem in general, and particularly in the medical domain. Currently, either trial-and-error methodologies or sampling-based ones are considered. This paper describes some preliminary steps towards effectively facing this task. The first step consists in the use of Differential Evolution, a kind of an Evolutionary Algorithm. The second lies in using a parallelized version in order to reduce the turnaround time. The preliminary results obtained here show that this approach can be useful in easily obtaining structures that allow increases in the network accuracy with respect to those provided by humans.

Pacienti s diabetem jsou vystaveni několika vážným zdravotním rizikům, které mohou vést k smrtelným následkům. Pro zvýšení kvality života těchto pacientů je nutné navrhnout umělou slinivku, která bude podávat inzulín v případě potřeby. Tento článek představuje algoritmus založený na genetickém programování pro extrapolaci regresního modelu schopného odhadnout hladinu glukózy v krvi (BG) pomocí měření intersticiální glukózy (IG) a jejich derivácí. Tento algoritmus představuje možný krok při budování základního prvku umělé slinivky, jmenovitě nové evoluční výpočetní metodologie pro odvození matematického vztahu mezi BG a IG. Navržený evoluční postup je vyhodnocen na základě reálné databáze, která se skládá z měření BG i IG u osob trpících diabetem 1. typu. Zjištěný model je validován porovnáním s dalšími technikami během experimentální fáze. The patients suffering from diabetes are subjected to several serious medical risks that can lead also to fatal consequences. To enhance the quality of life of these patients there is the necessity to devise an artificial pancreas able to inject an insulin bolus when needed. This paper presents a genetic-programming based algorithm to extrapolate a regression model able to estimate the blood glucose (BG) level through interstitial glucose (IG) measurements and their derivatives. This algorithm represents a possible step in building the fundamental element of such an artificial pancreas, namely a new evolutionary computation-based metodology to derive a mathematical relationship between BG and IG. The proposed evolutionary automatic procedure is evaluated on a real-world database made up of both BG and IG measurements of people suffering from Type 1 diabetes. The discovered model is validated through a comparison with other techniques during the experimental phase

The wide availability of sensing devices in the medical domain causes the creation of large and very large data sets. Hence, tasks as the classification in such data sets becomes more and more difficult. Deep Neural Networks (DNNs) are very effective in classification, yet finding the best values for their hyper-parameters is a difficult and time-consuming task. This paper introduces an approach to decrease execution times to automatically find good hyper-parameter values for DNN through Evolutionary Algorithms when classification task is faced. This decrease is obtained through the combination of two mechanisms. The former is constituted by a distributed version for a Differential Evolution algorithm. The latter is based on a procedure aimed at reducing the size of the training set and relying on a decomposition into cubes of the space of the data set attributes. Experiments are carried out on a medical data set about Obstructive Sleep Anpnea. They show that sub-optimal DNN hyper-parameter values are obtained in a much lower time with respect to the case where this reduction is not effected, and that this does not come to the detriment of the accuracy in the classification over the test set items.

The diabetes is correlated to a malfunction of the pancreas that produces very little or no insulin. A way to improve the quality of life of people with diabetes is to implement an artificial pancreas able to inject an insulin bolus when necessary. The aim of this paper is to devise a possibly step in constructing the fundamental element of such an artificial pancreas - estimation of the blood glucose (BG) through interstitial glucose (IG) measurements. In particular, a new methodology is presented to derive a mathematical relationship between BG and IG by exploiting the ability of the evolutionary techniques in solving this regression task. An automatic procedure is used to estimate the missing BG values within this database. To validate the discovered model a comparison with other models is carried out during the experimental phase.

In this paper an automatic method for the selection of those Fourier descriptors which better correlate a 2D shape contour is presented. To this aim, shape description has been modeled as a non linear approximation problem and a strict relationship between transform entropy and the sorted version of the transformed analysed boundary is derived. As a result, Fourier descriptors are selected in a hierarchical way and the minimum number of coefficients able to give a nearly optimal shape boundary representation is automatically derived. The latter maximizes an entropic interpretation of a complexity-based similarity measure, i.e. the normalized information distance. Preliminary experimental results show that the proposed method is able to provide a compact and computationally effective description of shape boundary which guarantees a nearly optimal matching with the original one.

The purpose of this paper is to conduct a systematic review of cyberbullying intervention programs that are either in current practice and/or have been well documented. There are negative effects associated with cyberbullying including, poor academic, social, and mental health outcomes. Consequently, there is a need to develop evidence-based interventions. Critical content and evaluation elements of traditional bullying interventions provided a framework, due to the limited evidence on effective features for addressing cyberbullying. The review is based on a set of criteria for traditional bullying by Craig, Pepler, and Shelley (2004), which emphasizes scientific merit and ease of implementation. Results suggest that most studies are lacking in scientific merit, with most studies meeting less than half of the criteria. The average ease of implementation score was higher than that of scientific merit, however only 3 programs provided maintenance after implementation. Recommendations are made for best practices for cyberbullying interventions.