Adverse drug reaction classification with deep neural networks

We study the problem of detecting sentences describing adverse drug reactions (ADRs) and frame the problem as binary classification. We investigate different neural network (NN) architectures for ADR classification. In particular, we propose two new neural network models, Convolutional Recurrent Neural Network (CRNN) by concatenating convolutional neural networks with recurrent neural networks, and Convolutional Neural Network with Attention (CNNA) by adding attention weights into convolutional neural networks. We evaluate various NN architectures on a Twitter dataset containing informal language and an Adverse Drug Effects (ADE) dataset constructed by sampling from MEDLINE case reports. Experimental results show that all the NN architectures outperform the traditional maximum entropy classifiers trained from n-grams with different weighting strategies considerably on both datasets. On the Twitter dataset, all the NN architectures perform similarly. But on the ADE dataset, CNN performs better than other more complex CNN variants. Nevertheless, CNNA allows the visualisation of attention weights of words when making classification decisions and hence is more appropriate for the extraction of word subsequences describing ADRs.

Detection and Grading of Prostate Cancer in Temporal Enhanced Ultrasound using Convolutional Neural Networks

Prostate cancer is the most common non-dermatological cancer amongst men in the developed world. The current definitive diagnosis is core needle biopsy guided by transrectal ultrasound. However, this method suffers from low sensitivity and specificity in detecting cancer. Recently, a new ultrasound based tissue typing approach has been proposed, known as temporal enhanced ultrasound (TeUS). In this approach, a set of temporal ultrasound frames is collected from a stationary tissue location without any intentional mechanical excitation. The main aim of this thesis is to implement a deep learning-based solution for prostate cancer detection and grading using TeUS data. In the proposed solution, convolutional neural networks are trained to extract high-level features from time domain TeUS data in temporally and spatially adjacent frames in nine in vivo prostatectomy cases. This approach avoids information loss due to feature extraction and also improves cancer detection rate. The output likelihoods of two TeUS arrangements are then combined to form our novel decision support system. This deep learning-based approach results in the area under the receiver operating characteristic curve (AUC) of 0.80 and 0.73 for prostate cancer detection and grading, respectively, in leave-one-patient-out cross-validation. Recently, multi-parametric magnetic resonance imaging (mp-MRI) has been utilized to improve detection rate of aggressive prostate cancer. In this thesis, for the first time, we present the fusion of mp-MRI and TeUS for characterization of prostate cancer to compensates the deficiencies of each image modalities and improve cancer detection rate. The results obtained using TeUS are fused with those attained using consolidated mp-MRI maps from multiple MR modalities and cancer delineations on those by multiple clinicians. The proposed fusion approach yields the AUC of 0.86 in prostate cancer detection. The outcomes of this thesis emphasize the viable potential of TeUS as a tissue typing method. Employing this ultrasound-based intervention, which is non-invasive and inexpensive, can be a valuable and practical addition to enhance the current prostate cancer detection.

Optimal control of non-linear systems through hybrid cell-mapping/artificial-neural-networks techniques

In this paper, a real-time optimal control technique for non-linear plants is proposed. The control system makes use of the cell-mapping (CM) techniques, widely used for the global analysis of highly non-linear systems. The CM framework is employed for designing approximate optimal controllers via a control variable discretization. Furthermore, CM-based designs can be improved by the use of supervised feedforward artificial neural networks (ANNs), which have proved to be universal and efficient tools for function approximation, providing also very fast responses. The quantitative nature of the approximate CM solutions fits very well with ANNs characteristics. Here, we propose several control architectures which combine, in a different manner, supervised neural networks and CM control algorithms. On the one hand, different CM control laws computed for various target objectives can be employed for training a neural network, explicitly including the target information in the input vectors. This way, tracking problems, in addition to regulation ones, can be addressed in a fast and unified manner, obtaining smooth, averaged and global feedback control laws. On the other hand, adjoining CM and ANNs are also combined into a hybrid architecture to address problems where accuracy and real-time response are critical. Finally, some optimal control problems are solved with the proposed CM, neural and hybrid techniques, illustrating their good performance.

Digital soil mapping using reference area and artificial neural networks.

Digital soil mapping is an alternative for the recognition of soil classes in areas where pedological surveys are not available. The main aim of this study was to obtain a digital soil map using artificial neural networks (ANN) and environmental variables that express soillandscape relationships. This study was carried out in an area of 11,072 ha located in the Barra Bonita municipality, state of São Paulo, Brazil. A soil survey was obtained from a reference area of approximately 500 ha located in the center of the area studied. With the mapping units identified together with the environmental variables elevation, slope, slope plan, slope profile, convergence index, geology and geomorphic surfaces, a supervised classification by ANN was implemented. The neural network simulator used was the Java NNS with the learning algorithm "back propagation." Reference points were collected for evaluating the performance of the digital map produced. The occurrence of soils in the landscape obtained in the reference area was observed in the following digital classification: medium-textured soils at the highest positions of the landscape, originating from sandstone, and clayey loam soils in the end thirds of the hillsides due to the greater presence of basalt. The variables elevation and slope were the most important factors for discriminating soil class through the ANN. An accuracy level of 82% between the reference points and the digital classification was observed. The methodology proposed allowed for a preliminary soil classification of an area not previously mapped using mapping units obtained in a reference area

Exact Combinatorial Optimization with Graph Convolutional Neural Networks

Combinatorial optimization problems are typically tackled by the branch-and-bound paradigm. We propose to learn a variable selection policy for branch-and-bound in mixed-integer linear programming, by imitation learning on a diversified variant of the strong branching expert rule. We encode states as bipartite graphs and parameterize the policy as a graph convolutional neural network. Experiments on a series of synthetic problems demonstrate that our approach produces policies that can improve upon expert-designed branching rules on large problems, and generalize to instances significantly larger than seen during training.

Interpretable Convolutional Neural Networks for Decoding and Analyzing Neural Time Series Data

Machine learning is widely adopted to decode multi-variate neural time series, including electroencephalographic (EEG) and single-cell recordings. Recent solutions based on deep learning (DL) outperformed traditional decoders by automatically extracting relevant discriminative features from raw or minimally pre-processed signals. Convolutional Neural Networks (CNNs) have been successfully applied to EEG and are the most common DL-based EEG decoders in the state-of-the-art (SOA). However, the current research is affected by some limitations. SOA CNNs for EEG decoding usually exploit deep and heavy structures with the risk of overfitting small datasets, and architectures are often defined empirically. Furthermore, CNNs are mainly validated by designing within-subject decoders. Crucially, the automatically learned features mainly remain unexplored; conversely, interpreting these features may be of great value to use decoders also as analysis tools, highlighting neural signatures underlying the different decoded brain or behavioral states in a data-driven way. Lastly, SOA DL-based algorithms used to decode single-cell recordings rely on more complex, slower to train and less interpretable networks than CNNs, and the use of CNNs with these signals has not been investigated. This PhD research addresses the previous limitations, with reference to P300 and motor decoding from EEG, and motor decoding from single-neuron activity. CNNs were designed light, compact, and interpretable. Moreover, multiple training strategies were adopted, including transfer learning, which could reduce training times promoting the application of CNNs in practice. Furthermore, CNN-based EEG analyses were proposed to study neural features in the spatial, temporal and frequency domains, and proved to better highlight and enhance relevant neural features related to P300 and motor states than canonical EEG analyses. Remarkably, these analyses could be used, in perspective, to design novel EEG biomarkers for neurological or neurodevelopmental disorders. Lastly, CNNs were developed to decode single-neuron activity, providing a better compromise between performance and model complexity.

Preharvest factors affecting quality and antioxidant levels on Abate Fétel pears: study of superficial scald with a multivariate statistical approach and neural networks

There are only a few insights concerning the influence that agronomic and management variability may have on superficial scald (SS) in pears. Abate Fétel pears were picked during three seasons (2018, 2019 and 2020) from thirty commercial orchards in the Emilia Romagna region, Italy. Using a multivariate statistical approach, high heterogeneity between farms for SS development after cold storage with regular atmosphere was demonstrated. Indeed, some factors seem to affect SS in all growing seasons: high yields, soil texture, improper irrigation and Nitrogen management, use of plant growth regulators, late harvest, precipitations, Calcium and cow manure, presence of nets, orchard age, training system and rootstock. Afterwards, we explored the spatio/temporal variability of fruit attributes in two pear orchards. Environmental and physiological spatial variables were recorded by a portable RTK GPS. High spatial variability of the SS index was observed. Through a geostatistical approach, some characteristics, including soil electrical conductivity and fruit size, have been shown to be negatively correlated with SS. Moreover, regression tree analyses were applied suggesting the presence of threshold values of antioxidant capacity, total phenolic content, and acidity against SS. High pulp firmness and IAD values before storage, denoting a more immature fruit, appeared to be correlated with low SS. Finally, a convolution neural networks (CNN) was tested to detect SS and the starch pattern index (SPI) in pears for portable device applications. Preliminary statistics showed that the model for SS had low accuracy but good precision, and the CNN for SPI denoted good performances compared to the Ctifl and Laimburg scales. The major conclusion is that Abate Fétel pears can potentially be stored in different cold rooms, according to their origin and quality features, ensuring the best fruit quality for the final consumers. These results might lead to a substantial improvement in the Italian pear industry.

Predicting corticospinal excitability from pre-TMS EEG using convolutional neural networks

The amplitude of motor evoked potentials (MEPs) elicited by transcranial magnetic stimulation (TMS) of the primary motor cortex (M1) shows a large variability from trial to trial, although MEPs are evoked by the same repeated stimulus. A multitude of factors is believed to influence MEP amplitudes, such as cortical, spinal and motor excitability state. The goal of this work is to explore to which degree the variation in MEP amplitudes can be explained by the cortical state right before the stimulation. Specifically, we analyzed a dataset acquired on eleven healthy subjects comprising, for each subject, 840 single TMS pulses applied to the left M1 during acquisition of electroencephalography (EEG) and electromyography (EMG). An interpretable convolutional neural network, named SincEEGNet, was utilized to discriminate between low- and high-corticospinal excitability trials, defined according to the MEP amplitude, using in input the pre-TMS EEG. This data-driven approach enabled considering multiple brain locations and frequency bands without any a priori selection. Post-hoc interpretation techniques were adopted to enhance interpretation by identifying the more relevant EEG features for the classification. Results show that individualized classifiers successfully discriminated between low and high M1 excitability states in all participants. Outcomes of the interpretation methods suggest the importance of the electrodes situated over the TMS stimulation site, as well as the relevance of the temporal samples of the input EEG closer to the stimulation time. This novel decoding method allows causal investigation of the cortical excitability state, which may be relevant for personalizing and increasing the efficacy of therapeutic brain-state dependent brain stimulation (for example in patients affected by Parkinson’s disease).

Neural networks and statistical analysis for classification of corneal videokeratography maps based on Zernike coefficients: a quantitative comparison

PURPOSE: The main goal of this study was to develop and compare two different techniques for classification of specific types of corneal shapes when Zernike coefficients are used as inputs. A feed-forward artificial Neural Network (NN) and discriminant analysis (DA) techniques were used. METHODS: The inputs both for the NN and DA were the first 15 standard Zernike coefficients for 80 previously classified corneal elevation data files from an Eyesys System 2000 Videokeratograph (VK), installed at the Departamento de Oftalmologia of the Escola Paulista de Medicina, São Paulo. The NN had 5 output neurons which were associated with 5 typical corneal shapes: keratoconus, with-the-rule astigmatism, against-the-rule astigmatism, "regular" or "normal" shape and post-PRK. RESULTS: The NN and DA responses were statistically analyzed in terms of precision ([true positive+true negative]/total number of cases). Mean overall results for all cases for the NN and DA techniques were, respectively, 94% and 84.8%. CONCLUSION: Although we used a relatively small database, results obtained in the present study indicate that Zernike polynomials as descriptors of corneal shape may be a reliable parameter as input data for diagnostic automation of VK maps, using either NN or DA.

Transmission lines distance protection using artificial neural networks

This work presents the development and implementation of an artificial neural network based algorithm for transmission lines distance protection. This algorithm was developed to be used in any transmission line regardless of its configuration or voltage level. The described ANN-based algorithm does not need any topology adaptation or ANN parameters adjustment when applied to different electrical systems. This feature makes this solution unique since all ANN-based solutions presented until now were developed for particular transmission lines, which means that those solutions cannot be implemented in commercial relays. (c) 2011 Elsevier Ltd. All rights reserved.

Application of neural networks for pavement evaluation

As vias de comunicação são indispensáveis para o desenvolvimento de uma nação, económica e socialmente. Num mundo globalizado, onde tudo deve chegar ao seu destino no menor espaço de tempo, as vias de comunicação assumem um papel vital. Assim, torna-se essencial construir e manter uma rede de transportes eficiente. Apesar de não ser o método mais eficiente, o transporte rodoviário é muitas vezes o mais económico e possibilita o transporte porta-a-porta, sendo em muitos casos o único meio de transporte possível. Por estas razões, o modo rodoviário tem uma quota significativa no mercado dos transportes, seja de passageiros ou mercadorias, tornando-o extremamente importante na rede de transportes de um país. Os países europeus fizeram um grande investimento na criação de extensas redes de estradas, cobrindo quase todo o seu território. Neste momento, começa-se a atingir o ponto onde a principal preocu+ação das entidades gestoras de estradas deixa de ser a construção de novas vias, passando a focar-se na necessidade de manutenção e conservação das vias existentes. Os pavimentos rodoviários, como todas as outras construções, requerem manutenção de forma a garantir bons níveis de serviço com qualidade, conforto e segurança. Devido aos custos inerentes às operações de manutenção de pavimentos, estas devem rigorosamente e com base em critérios científicos bem definidos. Assim, pretende-se evitar intervenções desnecessárias, mas também impedir que os danos se tornem irreparáveis e economicamente prejudiciais, com repercussões na segurança dos utilizadores. Para se estimar a vida útil de um pavimento é essencial realizar primeiro a caracterização estrutural do mesmo. Para isso, torna-se necessário conhecer o tipo de estrutura de um pavimento, nomeadamente a espessura e o módulo de elasticidade constituintes. A utilização de métodos de ensaio não destrutivos é cada vez mais reconhecida como uma forma eficaz para obter informações sobre o comportamento estrutural de pavimentos. Para efectuar estes ensaios, existem vários equipamentos. No entanto, dois deles, o Deflectómetro de Impacto e o Radar de Prospecção, têm demonstrado ser particularmente eficientes para avaliação da capacidade de carga de um pavimento, sendo estes equipamentos utilizados no âmbito deste estudo. Assim, para realização de ensaios de carga em pavimentos, o equipamento Deflectómetro de Impacto tem sido utilizado com sucesso para medir as deflexões à superfície de um pavimento em pontos pré-determinados quando sujeito a uma carga normalizada de forma a simular o efeito da passagem da roda de um camião. Complementarmente, para a obtenção de informações contínuas sobre a estrutura de um pavimento, o equipamento Radar de Prospecção permite conhecer o número de camadas e as suas espessuras através da utilização de ondas electromagnéticas. Os dados proporcionam, quando usados em conjunto com a realização de sondagens à rotação e poços em alguns locais, permitem uma caracterização mais precisa da condição estrutural de um pavimento e o estabelecimento de modelos de resposta, no caso de pavimentos existentes. Por outro lado, o processamento dos dados obtidos durante os ensaios “in situ” revela-se uma tarefa morosa e complexa. Actualmente, utilizando as espessuras das camadas do pavimento, os módulos de elasticidade das camadas são calculados através da “retro-análise” da bacia de deflexões medida nos ensaios de carga. Este método é iterativo, sendo que um engenheiro experiente testa várias estruturas diferentes de pavimento, até se obter uma estrutura cuja resposta seja o mais próximo possível da obtida durante os ensaios “in Situ”. Esta tarefa revela-se muito dependente da experiência do engenheiro, uma vez que as estruturas de pavimento a serem testadas maioritariamente do seu raciocínio. Outra desvantagem deste método é o facto de apresentar soluções múltiplas, dado que diferentes estruturas podem apresentar modelos de resposta iguais. A solução aceite é, muitas vezes, a que se julga mais provável, baseando-se novamente no raciocínio e experiência do engenheiro. A solução para o problema da enorme quantidade de dados a processar e das múltiplas soluções possíveis poderá ser a utilização de Redes Neuronais Artificiais (RNA) para auxiliar esta tarefa. As redes neuronais são elementos computacionais virtuais, cujo funcionamento é inspirado na forma como os sistemas nervosos biológicos, como o cérebro, processam a informação. Estes elementos são compostos por uma série de camadas, que por sua vez são compostas por neurónios. Durante a transmissão da informação entre neurónios, esta é modificada pela aplicação de um coeficiente, denominado “peso”. As redes neuronais apresentam uma habilidade muito útil, uma vez que são capazes de mapear uma função sem conhecer a sua fórmula matemática. Esta habilidade é utilizada em vários campos científicos como o reconhecimento de padrões, classificação ou compactação de dados. De forma a possibilitar o uso desta característica, a rede deverá ser devidamente “treinada” antes, processo realizado através da introdução de dois conjuntos de dados: os valores de entrada e os valores de saída pretendidos. Através de um processo cíclico de propagação da informação através das ligações entre neurónios, as redes ajustam-se gradualmente, apresentando melhores resultados. Apesar de existirem vários tipos de redes, as que aparentam ser as mais aptas para esta tarefa são as redes de retro-propagação. Estas possuem uma característica importante, nomeadamente o treino denominado “treino supervisionado”. Devido a este método de treino, as redes funcionam dentro da gama de variação dos dados fornecidos para o “treino” e, consequentemente, os resultados calculados também se encontram dentro da mesma gama, impedindo o aparecimento de soluções matemáticas com impossibilidade prática. De forma a tornar esta tarefa ainda mais simples, foi desenvolvido um programa de computador, NNPav, utilizando as RNA como parte integrante do seu processo de cálculo. O objectivo é tornar o processo de “retro-análise” totalmente automático e prevenir erros induzidos pela falta de experiência do utilizador. De forma a expandir ainda mais as funcionalidades do programa, foi implementado um processo de cálculo que realiza uma estimativa da capacidade de carga e da vida útil restante do pavimento, recorrendo a dois critérios de ruína. Estes critérios são normalmente utilizados no dimensionamento de pavimentos, de forma a prevenir o fendilhamento por fadiga e as deformações permanentes. Desta forma, o programa criado permite a estimativa da vida útil restante de um pavimento de forma eficiente, directamente a partir das deflexões e espessuras das camadas, medidas nos ensaios “in situ”. Todos os passos da caracterização estrutural do pavimento são efectuados pelo NNPav, seja recorrendo à utilização de redes neuronais ou a processos de cálculo matemático, incluindo a correcção do módulo de elasticidade da camada de misturas betuminosas para a temperatura de projecto e considerando as características de tráfego e taxas de crescimento do mesmo. Os testes efectuados às redes neuronais revelaram que foram alcançados resultados satisfatórios. Os níveis de erros na utilização de redes neuronais são semelhantes aos obtidos usando modelos de camadas linear-elásticas, excepto para o cálculo da vida útil com base num dos critérios, onde os erros obtidos foram mais altos. No entanto, este processo revela-se bastante mais rápido e possibilita o processamento dos dados por pessoal com menos experiência. Ao mesmo tempo, foi assegurado que nos ficheiros de resultados é possível analisar todos os dados calculados pelo programa, em várias fases de processamento de forma a permitir a análise detalhada dos mesmos. A possibilidade de estimar a capacidade de carga e a vida útil restante de um pavimento, contempladas no programa desenvolvido, representam também ferramentas importantes. Basicamente, o NNPav permite uma análise estrutural completa de um pavimento, estimando a sua vida útil com base nos ensaios de campo realizados pelo Deflectómetro de Impacto e pelo Radar de Prospecção, num único passo. Complementarmente, foi ainda desenvolvido e implementado no NNPav um módulo destinado ao dimensionamento de pavimentos novos. Este módulo permite que, dado um conjunto de estruturas de pavimento possíveis, seja estimada a capacidade de carga e a vida útil daquele pavimento. Este facto permite a análise de uma grande quantidade de estruturas de pavimento, e a fácil comparação dos resultados no ficheiro exportado. Apesar dos resultados obtidos neste trabalho serem bastante satisfatórios, os desenvolvimentos futuros na aplicação de Redes Neuronais na avaliação de pavimentos são ainda mais promissores. Uma vez que este trabalho foi limitado a uma moldura temporal inerente a um trabalho académico, a possibilidade de melhorar ainda mais a resposta das RNA fica em aberto. Apesar dos vários testes realizados às redes, de forma a obter as arquitecturas que apresentassem melhores resultados, as arquitecturas possíveis são virtualmente ilimitadas e pode ser uma área a aprofundar. As funcionalidades implementadas no programa foram as possíveis, dentro da moldura temporal referida, mas existem muitas funcionalidades a serem adicinadas ou expandidas, aumentando a funcionalidade do programa e a sua produtividade. Uma vez que esta é uma ferramenta que pode ser aplicada ao nível de gestão de redes rodoviárias, seria necessário estudar e desenvolver redes similares de forma a avaliar outros tipos de estruturas de pavimentos. Como conclusão final, apesar dos vários aspectos que podem, e devem ser melhorados, o programa desenvolvido provou ser uma ferramenta bastante útil e eficiente na avaliação estrutural de pavimentos com base em métodos de ensaio não destrutivos.

FMRI 3D registration based on Fourier space subsets using neural networks

In this work, we present a neural network (NN) based method designed for 3D rigid-body registration of FMRI time series, which relies on a limited number of Fourier coefficients of the images to be aligned. These coefficients, which are comprised in a small cubic neighborhood located at the first octant of a 3D Fourier space (including the DC component), are then fed into six NN during the learning stage. Each NN yields the estimates of a registration parameter. The proposed method was assessed for 3D rigid-body transformations, using DC neighborhoods of different sizes. The mean absolute registration errors are of approximately 0.030 mm in translations and 0.030 deg in rotations, for the typical motion amplitudes encountered in FMRI studies. The construction of the training set and the learning stage are fast requiring, respectively, 90 s and 1 to 12 s, depending on the number of input and hidden units of the NN. We believe that NN-based approaches to the problem of FMRI registration can be of great interest in the future. For instance, NN relying on limited K-space data (possibly in navigation echoes) can be a valid solution to the problem of prospective (in frame) FMRI registration.