To see the other types of publications on this topic, follow the link: Combination of neural networks.
Dissertations / Theses on the topic 'Combination of neural networks'
Create a spot-on reference in APA, MLA, Chicago, Harvard, and other styles
Consult the top 50 dissertations / theses for your research on the topic 'Combination of neural networks.'
Next to every source in the list of references, there is an 'Add to bibliography' button. Press on it, and we will generate automatically the bibliographic reference to the chosen work in the citation style you need: APA, MLA, Harvard, Chicago, Vancouver, etc.
You can also download the full text of the academic publication as pdf and read online its abstract whenever available in the metadata.
Browse dissertations / theses on a wide variety of disciplines and organise your bibliography correctly.
1
Morabito, David L. "Statistical mechanics of neural networks and combinatorial opimization problems /." Online version of thesis, 1991. http://hdl.handle.net/1850/11089.
Full text
2
Korn, Stefan. "The combination of AI modelling techniques for the simulation of manufacturing processes." Thesis, Glasgow Caledonian University, 1998. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.263139.
Full text
3
Amanzadi, Amirhossein. "Predicting safe drug combinations with Graph Neural Networks (GNN)." Thesis, Uppsala universitet, Institutionen för farmaceutisk biovetenskap, 2021. http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-446691.
Full textAbstract:
Many people - especially during their elderly - consume multiple drugs for the treatment of complex or co-existing diseases. Identifying side effects caused by polypharmacy is crucial for reducing mortality and morbidity of the patients which will lead to improvement in their quality of life. Since there is immense space for possible drug combinations, it is infeasible to examine them entirely in the lab. In silico models can offer a convenient solution, however, due to the lack of a sufficient amount of homogenous data it is difficult to develop both reliable and scalable models in its ability to accurately predict Polypharmacy Side Effect. Recent advancement in the field of representational learning has utilized the power of graph networks to harmonize information from the heterogeneous biological databases and interactomes. This thesis takes advantage of those techniques and incorporates them with the state-of-the-art Graph Neural Network algorithms to implement a Deep learning pipeline capable of predicting the Adverse Drug Reaction of any given paired drug combinations.
4
Freitas, Paulo Sérgio Abreu. "The combination of neural estimates in prediction and decision problems." Doctoral thesis, Universidade de Lisboa: Faculdade de Ciências, 2008. http://hdl.handle.net/10400.13/98.
Full textAbstract:
In this dissertation, different ways of combining neural predictive models or neural-based forecasts are discussed. The proposed approaches consider mostly Gaussian radial basis function networks, which can be efficiently identified and estimated through
recursive/adaptive methods. Two different ways of combining are explored to get a final
estimate – model mixing and model synthesis –, with the aim of obtaining improvements
both in terms of efficiency and effectiveness. In the context of model mixing, the usual framework for linearly combining estimates from different models is extended, to deal with the case where the forecast errors from those models are correlated. In the context of
model synthesis, and to address the problems raised by heavily nonstationary time series,
we propose hybrid dynamic models for more advanced time series forecasting, composed
of a dynamic trend regressive model (or, even, a dynamic harmonic regressive model), and a Gaussian radial basis function network. Additionally, using the model mixing procedure, two approaches for decision-making from forecasting models are discussed and compared: either inferring decisions from combined predictive estimates, or combining prescriptive
solutions derived from different forecasting models. Finally, the application of some of the models and methods proposed previously is illustrated with two case studies, based on time series from finance and from tourism.Orientador: António José Lopes Rodrigues
5
Yang, Shuang. "Multistage neural network ensemble : adaptive combination of ensemble results." Thesis, London Metropolitan University, 2003. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.425920.
Full text
6
Torres, Sospedra Joaquín. "Ensembles of Artificial Neural Networks: Analysis and Development of Design Methods." Doctoral thesis, Universitat Jaume I, 2011. http://hdl.handle.net/10803/48638.
Full textAbstract:
This thesis is focused on the analysis and development of Ensembles of Neural Networks. An ensemble is a system in which a set of heterogeneous Artificial Neural Networks are generated in order to outperform the Single network based classifiers. However, this proposed thesis differs from others related to ensembles of neural networks [1, 2, 3, 4, 5, 6, 7] since it is organized as follows.
In this thesis, firstly, an ensemble methods comparison has been introduced in order to provide a rank-based list of the best ensemble methods existing in the bibliography. This comparison has been split into two researches which represents two chapters of the thesis.
Moreover, there is another important step related to the ensembles of neural networks which is how to combine the information provided by the neural networks in the ensemble. In the bibliography, there are some alternatives to apply in order to get an accurate combination of the information provided by the heterogeneous set of networks. For this reason, a combiner comparison has also been introduced in this thesis.
Furthermore, Ensembles of Neural Networks is only a kind of Multiple Classifier System based on neural networks. However, there are other alternatives to generate MCS based on neural networks which are quite different to Ensembles. The most important systems are Stacked Generalization and Mixture of Experts. These two systems will be also analysed in this thesis and new alternatives are proposed.
One of the results of the comparative research developed is a deep understanding of the field of ensembles. So new ensemble methods and combiners can be designed after analyzing the results provided by the research performed. Concretely, two new ensemble methods, a new ensemble methodology called Cross-Validated Boosting and two reordering algorithms are proposed in this thesis. The best overall results are obtained by the ensemble methods proposed.
Finally, all the experiments done have been carried out on a common experimental setup. The experiments have been repeated ten times on nineteen different datasets from the UCI repository in order to validate the results. Moreover, the procedure applied to set up specific parameters is quite similar in all the experiments performed.
It is important to conclude by remarking that the main contributions are:
1) An experimental setup to prepare the experiments which can be applied for further comparisons. 2) A guide to select the most appropriate methods to build and combine ensembles and multiple classifiers systems. 3) New methods proposed to build ensembles and other multiple classifier systems.
7
Henry, Timothy G. "Generalization of deep neural networks to unseen attribute combinations." Thesis, Massachusetts Institute of Technology, 2020. https://hdl.handle.net/1721.1/129905.
Full textAbstract:
Thesis: M. Eng., Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, February, 2020Cataloged from student-submitted PDF of thesis.
Includes bibliographical references (pages 71-73).
Visual understanding results from a combined understanding of primitive visual attributes such as color, texture, and shape. This allows humans and other primates to generalize their understanding of objects to new combinations of attributes. For instance, one can understand that a pink elephant is an elephant even if they have never seen this particular combination of color and shape before. However, is it the case that deep neural networks (DNNs) are able to generalize to such novel combinations in object recognition or other related vision tasks? This thesis demonstrates that (1) the ability of DNNs to generalize to unseen attribute combinations increases with the increased diversity of combinations seen in training as a percentage of the total combination space, (2) this effect is largely independent of the specifics of the DNN architecture used, (3) while single-task and multi-task formulations of supervised attribute classification problems may lead to similar performance on seen combinations, single-task formulations have a superior ability to generalize to unseen combinations, and (4) DNNs demonstrating the ability to generalize well in this setting learn to do so by leveraging emergent hidden units that exhibit properties of attribute selectivity and invariance.
by Timothy G. Henry.
M. Eng.
M.Eng. Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science
8
Zhao, Yi. "Combination of Wireless sensor network and artifical neuronal network : a new approach of modeling." Thesis, Toulon, 2013. http://www.theses.fr/2013TOUL0013/document.
Full textAbstract:
Face à la limitation de la modélisation paramétrique, nous avons proposé dans cette thèse une procédure standard pour combiner les données reçues a partir de Réseaux de capteurs sans fils (WSN) pour modéliser a l'aide de Réseaux de Neurones Artificiels (ANN). Des expériences sur la modélisation thermique ont permis de démontrer que la combinaison de WSN et d'ANN est capable de produire des modèles thermiques précis. Une nouvelle méthode de formation "Multi-Pattern Cross Training" (MPCT) a également été introduite dans ce travail. Cette méthode permet de fusionner les informations provenant de différentes sources de données d'entraînements indépendants (patterns) en un seul modèle ANN. D'autres expériences ont montré que les modèles formés par la méthode MPCT fournissent une meilleure performance de généralisation et que les erreurs de prévision sont réduites. De plus, le modèle de réseau neuronal basé sur la méthode MPCT a montré des avantages importants dans le multi-variable Model Prédictive Control (MPC). Les simulations numériques indiquent que le MPC basé sur le MPCT a surpassé le MPC multi-modèles au niveau de l'efficacité du contrôleA Wireless Sensor Network (WSN) consisting of autonomous sensor nodes can provide a rich stream of sensor data representing physical measurements. A well built Artificial Neural Network (ANN) model needs sufficient training data sources. Facing the limitation of traditional parametric modeling, this paper proposes a standard procedure of combining ANN and WSN sensor data in modeling. Experiments on indoor thermal modeling demonstrated that WSN together with ANN can lead to accurate fine grained indoor thermal models. A new training method "Multi-Pattern Cross Training" (MPCT) is also introduced in this work. This training method makes it possible to merge knowledge from different independent training data sources (patterns) into a single ANN model. Further experiments demonstrated that models trained by MPCT method shew better generalization performance and lower prediction errors in tests using different data sets. Also the MPCT based Neural Network Model has shown advantages in multi-variable Neural Network based Model Predictive Control (NNMPC). Software simulation and application results indicate that MPCT implemented NNMPC outperformed Multiple models based NNMPC in online control efficiency
9
Alani, Shayma. "Design of intelligent ensembled classifiers combination methods." Thesis, Brunel University, 2015. http://bura.brunel.ac.uk/handle/2438/12793.
Full textAbstract:
Classifier ensembling research has been one of the most active areas of machine learning for a long period of time. The main aim of generating combined classifier ensembles is to improve the prediction accuracy in comparison to using an individual classifier. A combined classifiers ensemble can improve the prediction results by compensating for the individual classifier weaknesses in certain areas and benefiting from better accuracy of the other ensembles in the same area. In this thesis, different algorithms are proposed for designing classifier ensemble combiners. The existing methods such as averaging, voting, weighted average, and optimised weighted method does not increase the accuracy of the combiner in comparison to the proposed advanced methods such as genetic programming and the coalition method. The different methods are studied in detail and analysed using different databases. The aim is to increase the accuracy of the combiner in comparison to the standard stand-alone classifiers. The proposed methods are based on generating a combiner formula using genetic programming, while the coalition is based on estimating the diversity of the classifiers such that a coalition is generated with better prediction accuracy. Standard accuracy measures are used, namely accuracy, sensitivity, specificity and area under the curve, in addition to training error accuracies such as the mean square error. The combiner methods are compared empirically with several stand-alone classifiers using neural network algorithms. Different types of neural network topologies are used to generate different models. Experimental results show that the combiner algorithms are superior in creating the most diverse and accurate classifier ensembles. Ensembles of the same models are generated to boost the accuracy of a single classifier type. An ensemble of 10 models of different initial weights is used to improve the accuracy. Experiments show a significant improvement over a single model classifier. Finally, two combining methods are studied, namely the genetic programming and coalition combination methods. The genetic programming algorithm is used to generate a formula for the classifiers’ combinations, while the coalition method is based on a simple algorithm that assigns linear combination weights based on the consensus theory. Experimental results of the same databases demonstrate the effectiveness of the proposed methods compared to conventional combining methods. The results show that the coalition method is better than genetic programming.
10
Huhtinen, J. (Jouni). "Utilization of neural network and agent technology combination for distributed intelligent applications and services." Doctoral thesis, University of Oulu, 2005. http://urn.fi/urn:isbn:9514278550.
Full textAbstract:
Abstract
The use of agent systems has increased enormously, especially in the field of mobile services. Intelligent services have also increased rapidly in the web. In this thesis, the utilization of software agent technology in mobile services and decentralized intelligent services in the multimedia business is introduced and described. Both Genie Agent Architecture (GAA) and Decentralized International and Intelligent Software Architecture (DIISA) are described.
The common problems in decentralized software systems are lack of intelligence, communication of software modules and system learning. Another problem is the personalization of users and services. A third problem is the matching of users and service characteristics in web application level in a non-linear way. In this case it means that web services follow human steps and are capable of learning from human inputs and their characteristics in an intelligent way. This third problem is addressed in this thesis and solutions are presented with two intelligent software architectures and services.
The solutions of the thesis are based on a combination of neural network and agent technology. To be more specific, solutions are based on an intelligent agent which uses certain black box information like Self-Organized Map (SOM). This process is as follows; information agents collect information from different sources like the web, databases, users, other software agents and the environment. Information is filtered and adapted for input vectors. Maps are created from a data entry of an SOM. Using maps is very simple, input forms are completed by users (automatically or manually) or user agents. Input vectors are formed again and sent to a certain map. The map gives several outputs which are passed through specific algorithms. This information is passed to an intelligent agent.
The needs for web intelligence and knowledge representation serving users is a current issue in many business solutions. The main goal is to enable this by means of autonomous agents which communicate with each other using an agent communication language and with users using their native languages via several communication channels.
11
Prampero, Paulo Sérgio. "Combinação de Classificadores para Reconhecimento de Padrões." Universidade de São Paulo, 1998. http://www.teses.usp.br/teses/disponiveis/55/55134/tde-16032018-090228/.
Full textAbstract:
O cérebro humano é formado por um conjunto de neurônios de diferentes tipos, cada um com sua especialidade. A combinação destes diferentes tipos de neurônios é um dos aspectos responsáveis pelo desempenho apresentado pelo cérebro na realização de várias tarefas. Redes Neurais Artificiais são técnicas computacionais que apresentam um modelo matemático inspirado no sistema nervoso e que adquirem conhecimento através da experiência. Uma alternativa para melhorar o desempenho das Redes Neurais Artificiais é a utilização de técnicas de Combinação de Classificadores. Estas técnicas de combinação exploram as diferenças e as semelhanças das redes para a obtenção de resultados melhores. Dentre as principais aplicações de Redes Neurais Artificiais está o Reconhecimento de Padrões. Neste trabalho, foram utilizadas técnicas de Combinação de Classificadores para a combinação de Redes Neurais Artificiais em problemas de Reconhecimento de Padrões.The human brain is formed by neurons of different types, each one with its own speciality. The combination of theses different types of neurons is one of the main features responsible for the brain performance in severa! tasks. Artificial Neural Networks are computation technics whose mathematical model is based on the nervous system and learns new knowledge by experience. An alternative to improve the performance of Artificial Neural Networks is the employment of Classifiers Combination techniques. These techniques of combination explore the difference and the similarity of the networks to achieve better performance. The main application of Artificial Neural Networks is Pattern Recognition. In this work, Classifiers Combination techniques were utilized to combine Artificial Neural Networks to solve Pattern Recognition problems.
12
Viola, Federica. "Automatic Sleep Scoring To Study Brain Resting State Networks During Sleep In Narcoleptic And Healthy Subjects : A Combination Of A Wavelet Filter Bank And An Artificial Neural Network." Thesis, Linköpings universitet, Institutionen för medicin och hälsa, 2014. http://urn.kb.se/resolve?urn=urn:nbn:se:liu:diva-110950.
Full textAbstract:
Manual sleep scoring, executed by visual inspection of the EEG, is a very time consuming activity, with an inherent subjective decisional component. Automatic sleep scoring could ease the job of the technicians, because faster and more accurate. Frequency information characterizing the main brain rhythms, and consequently the sleep stages, needs to be extracted from the EEG data. The approach used in this study involves a wavelet filter bank for the EEG frequency features extraction. The wavelet packet analysis tool in MATLAB has been employed and the frequency information subsequently used for the automatic sleep scoring by means of an artificial neural network. Finally, the automatic sleep scoring has been employed for epoching the fMRI data, thus allowing for studying brain resting state networks during sleep. Three resting state networks have been inspected; the Default Mode Network, The Attentional Network and the Salience Network. The networks functional connectivity variations have been inspected in both healthy and narcoleptic subjects. Narcolepsy is a neurobiological disorder characterized by an excessive daytime sleepiness, whose aetiology may be linked to a loss of neurons in the hypothalamic region.
13
Помпа, Костянтин Віталійович. "Нейронна мережа для виявлення повторних новоутворень у мозку пацієнта на МРТ-зображенні." Master's thesis, Київ, 2019. https://ela.kpi.ua/handle/123456789/27735.
Full textAbstract:
Обсяг магістерської дисертації становить 77 сторінок, містить 25 рисунків, 6 таблиць. Загалом було опрацьовано 53 джерела.
Робота присвячена створенню нейронної мережі для виявлення рецидивів пухлини головного мозку. Розроблену мережу можливо застосовувати для післяопераційного моніторингу змін в області, враженій пухлиною, а також для досліджень в області нейронних мереж та медицини.
Метою роботи є підвищення ефективності виявлення рецидивів пухлини мозку шляхом створення нейронної мережі для сегментації МРТ зображень.
Об’єктом дослідження є нейромережа для сегментації МРТ зображень.
Предметом дослідження виступає алгоритм сегментації пухлини головного мозку за допомогою нейронної мережі на зображеннях МРТ.
У магістерській дисертації обґрунтовано необхідність створення інформаційної нейронної мережі, її ефективність порівняно з іншими існуючими нейромережами. Розроблена нейронна мережа дозволяє виявити рецидиви пухлин головного мозку. До того ж, архітектуру даної нейромережевої комбінації можна вдосконалити.
В середовищі розробки Python створено нейромережевий ансамбль та перевірено точність розпізнавання ним пухлин головного мозку.The volume of the master's dissertation is 77 pages, contains 25 figures, 6 tables. In total, 53 sources were processed. The work is devoted to the creation of an informative neural network for the detection of recurrence of the brain tumor. The developed system can be used for postoperative monitoring of changes in the affected area of the tumor, as well as for research in the field of neural networks and medicine. The purpose of the work is to create an informative neural network for the detection of recurrent neoplasms in the patient's brain. The object of the study is a neural network for the segmentation of MRI images. The subject of the study is the characteristics of MRI images and the information neural network tested in the development environment of Python. In the master's dissertation the necessity of creation of the informative neural network, its efficiency in comparison with other existing neural networks is substantiated. The developed neural network allows to detect recurrence of tumors of the brain. In addition, the architecture of this neural network combination can be improved. In the development environment of Python, a neural network ensemble was created and the accuracy of the recognition of the brain tumors was checked.
14
Saragiotis, Panagiotis. "Cross-modal classification and retrieval of multimodal data using combinations of neural networks." Thesis, University of Surrey, 2006. http://epubs.surrey.ac.uk/843338/.
Full textAbstract:
Current neurobiological thinking supported, in part, by experimentation stresses the importance of cross-modality. Uni-modal cognitive tasks, language and vision, for example, are performed with the help of many networks working simultaneously or sequentially; and for cross-modal tasks, like picture / object naming and word illustration, the output of these networks is combined to produce higher cognitive behaviour. The notion of multi-net processing is used typically in the pattern recognition literature, where ensemble networks of weak classifiers - typically supervised - appear to outperform strong classifiers. We have built a system, based on combinations of neural networks, that demonstrates how cross-modal classification can be used to retrieve multi-modal data using one of the available modalities of information. Two multi-net systems were used in this work: one comprising Kohonen SOMs that interact with each other via a Hebbian network and a fuzzy ARTMAP network where the interaction is through the embedded map field. The multi-nets were used for the cross-modal retrieval of images given keywords and for finding the appropriate keywords for an image. The systems were trained on two publicly available image databases that had collateral annotations on the images. The Hemera collection, comprising images of pre-segmented single objects, and the Corel collection with images of multiple objects were used for automatically generating various sets of input vectors. We have attempted to develop a method for evaluating the performance of multi-net systems using a monolithic network trained on modally-undifferentiated vectors as an intuitive bench-mark. To this extent single SOM and fuzzy ART networks were trained using a concatenated visual / linguistic vector to test the performance of multi-net systems with typical monolithic systems. Both multi-nets outperform the respective monolithic systems in terms of information retrieval measures of precision and recall on test images drawn from both datasets; the SOM multi-net outperforms the fuzzy ARTMAP both in terms of convergence and precision-recall. The performance of the SOM-based multi-net in retrieval, classification and auto-annotation is on a par with that of state of the art systems like "ALIP" and "Blobworld". Much of the neural network based simulations reported in the literature use supervised learning algorithms. Such algorithms are suited when classes of objects are predefined and objects in themselves are quite unique in terms of their attributes. We have compared the performance of our multi-net systems with that of a multi-layer perceptron (MLP). The MLP does show substantially greater precision and recall on a (fixed) class of objects when compared with our unsupervised systems. However when 'lesioned' -the network connectivity 'damaged' deliberately- the multi-net systems show a greater degree of robustness. Cross-modal systems appear to hold considerable intellectual and commercial potential and the multi-net approach facilitates the simulation of such systems.
15
Butler, Martin A. "A Method of Structural Health Monitoring for Unpredicted Combinations of Damage." University of Cincinnati / OhioLINK, 2019. http://rave.ohiolink.edu/etdc/view?acc_num=ucin1575967420002943.
Full text
16
Antoniou, Christos Andrea. "Improving the acoustic modelling of speech using modular/ensemble combinations of heterogeneous neural networks." Thesis, University of Essex, 2000. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.340582.
Full text
17
Coelho, Guilherme Palermo 1980. "Geração, seleção e combinação de componentes para ensembles de redes neurais aplicadas a problemas de classificação." [s.n.], 2006. http://repositorio.unicamp.br/jspui/handle/REPOSIP/261408.
Full textAbstract:
Orientador: Fernando Jose Von ZubenDissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Eletrica e Computação
Made available in DSpace on 2018-08-11T19:03:12Z (GMT). No. of bitstreams: 1 Coelho_GuilhermePalermo_M.pdf: 2968179 bytes, checksum: bbea7c9c565907f86eee09155421bfa3 (MD5) Previous issue date: 2006
Resumo: O uso da abordagem ensembles tem sido bastante explorado na última década, por se tratar de uma técnica simples e capaz de aumentar a capacidade de generalização de soluções baseadas em aprendizado de máquina. No entanto, para que um ensemble seja capaz de promover melhorias de desempenho, os seus componentes devem apresentar bons desempenhos individuais e, ao mesmo tempo, devem ter comportamentos diversos entre si. Neste trabalho, é proposta uma metodologia de criação de ensembles para problemas de classificação, onde os componentes são redes neurais artificiais do tipo perceptron multicamadas. Para que fossem gerados bons candidatos a comporem o ensemble, atendendo a critérios de desempenho e de diversidade, foi aplicada uma meta-heurística populacional imuno-inspirada, denominada opt-aiNet, a qual é caracterizada por definir automaticamente o número de indivíduos na população a cada iteração, promover diversidade e preservar ótimos locais ao longo da busca. Na etapa de seleção dos componentes que efetivamente irão compor o ensemble, foram utilizadas seis técnicas distintas e, para combinação dos componentes selecionados, foram adotadas cinco estratégias. A abordagem proposta foi aplicada a quatro problemas de classificação de padrões e os resultados obtidos indicam a validade da metodologia de criação de ensembles. Além disso, foi verificada uma dependência entre o melhor par de técnicas de seleção e combinação e a população de indivíduos candidatos a comporem o ensemble, assim como foi feita uma análise de confiabilidade dos resultados de classificação
Abstract: In the last decade, the ensemble approach has been widely explored, once it is a simple technique capable of increasing the generalization capability of machine learning based solutions. However, an ensemble can only promote performance enhancement if its components present good individual performance and, at the same time, diverse behavior among each other. This work proposes a methodology to synthesize ensembles for classification problems, where the components of the ensembles are multi-layer perceptrons. To generate good candidates to compose the ensemble, meeting the performance and diversity requirements, it was applied a populational and immune-inspired metaheuristic, named opt-aiNet, which is characterized as being capable of automatically determining the number of individuals in the population at each iteration, promoting diversity and preserving local optima through the search. In the component selection phase, six distinct techniques were applied and, to combine these selected components, five strategies were adopted. The proposed approach was applied to four pattern classification problems and the obtained results indicated the validity of the methodology to synthesize ensembles. It was also verified a dependence of the best pair of selection and combination techniques on the population of candidates to compose the ensemble, and it was made an analysis of the confidence of the classification results
Mestrado
Engenharia de Computação
Mestre em Engenharia Elétrica
18
Valmiki, Geetha Charan, and Akhil Santosh Tirupathi. "Performance Analysis Between Combinations of Optimization Algorithms and Activation Functions used in Multi-Layer Perceptron Neural Networks." Thesis, Blekinge Tekniska Högskola, Institutionen för datavetenskap, 2020. http://urn.kb.se/resolve?urn=urn:nbn:se:bth-20204.
Full textAbstract:
Background:- Artificial Neural networks are motivated from biological nervous system and can be used for classification and forecasting the data. Each neural node contains activation function could be used for solving non-linear problems and optimization function to minimize the loss and give more accurate results. Neural networks are bustling in the field of machine learning, which inspired this study to analyse the performance variation based on the use of different combinations of the activation functions and optimization algorithms in terms of accuracy results and metrics recall and impact of data-set features on the performance of the neural networks. Objectives:- This study deals with an experiment to analyse the performance of the combinations are performing well and giving more results and to see impact of the feature segregation from data-set on the neural networks model performance. Methods:- The process involve the gathering of the data-sets, activation functions and optimization algorithm. Execute the network model using 7X5 different combinations of activation functions and optimization algorithm and analyse the performance of the neural networks. These models are tested upon the same data-set with some of the discarded features to know the effect on the performance of the neural networks. Results:- All the metrics for evaluating the neural networks presented in separate table and graphs are used to show growth and fall down of the activation function when associating with different optimization function. Impact of the individual feature on the performance of the neural network is also represented. Conclusions:- Out of 35 combinations, combinations made from optimizations algorithms Adam,RMSprop and Adagrad and activation functions ReLU,Softplus,Tanh Sigmoid and Hard_Sigmoid are selected based on the performance evaluation and data has impact on the performance of the combinations of the algorithms and activation functions which is also evaluated based on the experimentation. Individual features have their corresponding effect on the neural network.
19
Tian, Tian. "Domain Adaptation and Model Combination for the Annotation of Multi-source, Multi-domain Texts." Thesis, Paris 3, 2019. http://www.theses.fr/2019PA030003.
Full textAbstract:
Internet propose aujourd’hui aux utilisateurs de services en ligne de commenter, d’éditer et de partager leurs points de vue sur différents sujets de discussion. Ce type de contenu est maintenant devenu la ressource principale pour les analyses d’opinions sur Internet. Néanmoins, à cause des abréviations, du bruit, des fautes d’orthographe et toutes autres sortes de problèmes, les outils de traitements automatiques des langues, y compris les reconnaisseurs d’entités nommées et les étiqueteurs automatiques morphosyntaxiques, ont des performances plus faibles que sur les textes bien-formés (Ritter et al., 2011).Cette thèse a pour objet la reconnaissance d’entités nommées sur les contenus générés par les utilisateurs sur Internet. Nous avons établi un corpus d’évaluation avec des textes multi-sources et multi-domaines. Ensuite, nous avons développé un modèle de champs conditionnels aléatoires, entrainé sur un corpus annoté provenant des contenus générés par les utilisateurs.Dans le but d’améliorer les résultats de la reconnaissance d’entités nommées, nous avons d’abord développé un étiqueteur morpho-syntaxique sur les contenus générés par les utilisateurs et nous avons utilisé les étiquettesprédites comme un attribut du modèle des champs conditionnels aléatoire. Enfin, pour transformer les contenus générés par les utilisateurs en textes bien-formés, nous avons développé un modèle de normalisation lexicale basé sur des réseaux de neurones pour proposer une forme correcte pour les mots non-standardThe increasing mass of User-Generated Content (UGC) on the Internet means that people are now willing to comment, edit or share their opinions on different topics. This content is now the main ressource for sentiment analysis on the Internet. Due to abbreviations, noise, spelling errors and all other problems with UGC, traditional Natural Language Processing (NLP) tools, including Named Entity Recognizers and part-of-speech (POS) taggers, perform poorly when compared to their usual results on canonical text (Ritter et al., 2011).This thesis deals with Named Entity Recognition (NER) on some User-Generated Content (UGC). We have created an evaluation dataset including multi-domain and multi-sources texts. We then developed a Conditional Random Fields (CRFs) model trained on User-Generated Content (UGC).In order to improve NER results in this context, we first developed a POStagger on UGC and used the predicted POS tags as a feature in the CRFs model. To turn UGC into canonical text, we also developed a normalization model using neural networks to propose a correct form for Non-Standard Words (NSW) in the UGC
各种社交网络应用使得互联网用户对各种话题的实时评价，编辑和分享成为可能。这类用户生成的文本内容(User Generated content)已成为社交网络上意见分析的主要目标和来源。但是，此类文本内容中包含的缩写，噪声（不规则词），拼写错误以及其他各种问题导致包括命名实体识别，词性标注在内的传统的自然语言处理工具的性能，相比良好组成的文本降低了许多【参见Ritter 2011】。本论文的主要目标是针对社交网络上用户生成文本内容的命名实体识别。我们首先建立了一个包含多来源，多领域文本的有标注的语料库作为标准评价语料库。然后，我们开发了一个由社交网络用户生成文本训练的基于条件随机场(Conditional Random Fields)的序列标注模型。基于改善这个命名实体识别模型的目的，我们又开发了另一个同样由社交网络用户生成内容训练的词性标注模型，并使用此模型预测的词性作为命名实体识别的条件随机场模型的特征。最后，为了将用户生成文本内容转换成相对标准的良好文本内容，我们开发了一个基于神经网络的词汇标准化模型，用以改正用户生成文本内容中的不标准字，并使用模型提供的改正形式作为命名实体识别的条件随机场模型的特征，借以改善原模型的性能。
20
Jacobs, William. "COMBINAÇÃO DAS PREVISÕES DOS MODELOS DE BOX-JENKINS E MLP/RNA PARA A PREVISÃO DE DEMANDA NO PLANEJAMENTO DA PRODUÇÃO." Universidade Federal de Santa Maria, 2014. http://repositorio.ufsm.br/handle/1/8327.
Full textAbstract:
Coordenação de Aperfeiçoamento de Pessoal de Nível SuperiorA forecast of future demand for the products is the main variable to be considered in the planning and in production control in organizations. Two methods of time series forecasting often used in the literature are the ARIMA and MLP/RNA models. A practice that began in 1969 and has consolidated for greater accuracy is the combination of individual forecasts from two or more models. Considering the need for organizations by predictive techniques that generate better results, this study aims to predict the future values of a time series of the demand for UHT milk in a dairy industry, through the combination of ARIMA and MLP/RNA models, and to compare the results obtained by the combinations compared to individual models, exemplifying the achievement of combined forecasting in production planning. Accuracy measures to measure the results and to select the best model were the RMSE and MAPE for forecasting. The results showed that the combination of models SARIMA(3,0,1)(1,1,0)12 and DMLP the inverse mean square method provided a performance forecast for the six months ahead, up to 66.5% higher than individual models used, where the combination of the predictions obtained a RMSE of 1.43, and a MAPE of 2.16. In the 12 month ahead prediction for the performance of the combination was up to 56.5% higher compared to individual models, in which case obtained a RMSE of 2.86 and 3.70% MAPE. The combination of time series models enabled a significant increase in performance prediction models, but in order to produce satisfactory absolute results should be used to complement their predictive abilities mutually.
A previsão da demanda futura dos produtos é a principal variável a ser considerada no planejamento e controle da produção nas organizações. As técnicas de previsão de demanda são fundamentais no planejamento da produção de nível tático e operacional, especialmente as séries temporais, pois não requerem do planejador, uma investigação mais aprofundada acerca dos fatores que influenciam a demanda. Dois métodos de previsão de séries temporais frequentemente utilizados na literatura são os modelos ARIMA e os modelos MLP/RNA. Uma prática que surgiu em 1969 e já consolidada para obter maior acurácia é a combinação das previsões individuais de dois ou mais modelos. Considerando a necessidade das organizações por técnicas preditivas que gerem melhores resultados, este estudo tem como objetivo prever os valores futuros de uma série temporal da demanda de leite UHT em uma indústria de lácteos, por meio da combinação dos modelos ARIMA e MLP/RNA, e comparar os resultados obtidos pelas combinações em relação aos modelos individuais, exemplificando a obtenção da previsão combinada no planejamento da produção. As medidas de acurácia para mensurar os resultados obtidos e selecionar o melhor modelo, foram o RMSE e o MAPE de previsão. Os resultados mostraram que a combinação dos modelos SARIMA(3,0,1)(1,1,0)12 e DMLP pelo método inverse mean square forneceu um desempenho na previsão para 6 meses adiante, de até 66,5% superior em relação aos modelos individuais utilizados, onde a combinação das previsões obteve um RMSE de 1,43 e um MAPE de 2,16. Na previsão para 12 meses adiante, o desempenho da combinação foi de até 56,5% superior em relação aos modelos individuais, caso em que obteve um RMSE de 2,86 e um MAPE de 3,70%. A combinação de modelos de séries temporais possibilitou um aumento significativo no desempenho de previsão dos modelos, mas para que se obtenham resultados absolutos satisfatórios, devem-se utilizar modelos previsores que complementem mutuamente a capacidade preditiva.
21
Defilippo, Samuel Belini. "Previsão da demanda de energia elétrica por combinações de modelos lineares e de inteligência computacional." Universidade Federal de Juiz de Fora (UFJF), 2017. https://repositorio.ufjf.br/jspui/handle/ufjf/6036.
Full textAbstract:
Submitted by Geandra Rodrigues (geandrar@gmail.com) on 2018-01-17T11:13:15Z
No. of bitstreams: 1
samuelbelinidefilippo.pdf: 2610291 bytes, checksum: 6c4f48d00a0649b56977f6c8a7ada4e0 (MD5)Approved for entry into archive by Adriana Oliveira (adriana.oliveira@ufjf.edu.br) on 2018-01-22T16:33:53Z (GMT) No. of bitstreams: 1 samuelbelinidefilippo.pdf: 2610291 bytes, checksum: 6c4f48d00a0649b56977f6c8a7ada4e0 (MD5)
Made available in DSpace on 2018-01-22T16:33:53Z (GMT). No. of bitstreams: 1 samuelbelinidefilippo.pdf: 2610291 bytes, checksum: 6c4f48d00a0649b56977f6c8a7ada4e0 (MD5) Previous issue date: 2017-09-20
Todo a produção, transmissão e distribuição de energia elétrica ocorre concomitantemente com o consumo da energia. Isso é necessário porque ainda não existe hoje uma maneira viável de se estocar energia em grandes quantidades. Dessa forma, a energia gerada precisa ser consumida quase que instantaneamente. Isso faz com que as previsões de demanda sejam fundamentais para uma boa gestão dos sistemas de energia. Esse trabalho focaliza métodos de previsão de demanda a curto prazo, até um dia à frente. Nos métodos mais simples, as previsões são feitas por modelos lineares que utilizam dados históricos da demanda de energia. Contudo, modelos baseados em inteligência computacional têm sido estudados para este fim, por explorarem a relação não-linear entre a demanda de energia e as variáveis climáticas. Em geral, estes modelos conseguem melhores previsões do que os métodos lineares. Seus resultados, porém, são instáveis e sensíveis a erros de medição, gerando erros de previsão discrepantes, que podem ter graves consequências para o processo de produção. Neste estudo, empregamos redes neurais artificiais e algoritmos genéticos para modelar dados históricos de carga e de clima, e combinamos estes modelos com métodos lineares tradicionais. O objetivo é conseguir previsões que não apenas sejam mais acuradas em termos médios, mas que também menos sensíveis aos erros de medição.
The production, transmission and distribution of electric energy occurs concomitantly with its consumption. This is necessary because there is yet no feasible way to store energy in large quantities. Therefore, the energy generated must be consumed almost instantaneously. This makes forecasting essential for the proper management of energy systems. This thesis focuses on short-term demand forecasting methods up to one day ahead. In simpler methods, the forecasts are made by linear models, which use of historical data on energy demand. However, computer intelligence-based models have been studied for this end, exploring the nonlinear relationship between energy demand and climatic variables. In general, these models achieve better forecasts than linear methods. Their results, however, are unstable and sensitive to measurement errors, leading to outliers in forecasting errors, which can have serious consequences for the production process. In this thesis, we use artificial neural networks and genetic algorithms for modelling historical load and climate data, and combined these models with traditional linear methods. The aim is to achieve forecasts that are not only more accurate in mean terms, but also less sensitive to measurement errors.
22
Kennedy, Brian Alexander. "Developing the Cis-Regulatory Association Model (CRAM) to Identify Combinations of Transcription Factors in ChIP-Seq Data." The Ohio State University, 2010. http://rave.ohiolink.edu/etdc/view?acc_num=osu1291138540.
Full text
23
Stuner, Bruno. "Cohorte de réseaux de neurones récurrents pour la reconnaissance de l'écriture." Thesis, Normandie, 2018. http://www.theses.fr/2018NORMR024.
Full textAbstract:
Les méthodes à l’état de l’art de la reconnaissance de l’écriture sont fondées sur des réseaux de neurones récurrents (RNN) à cellules LSTM ayant des performances remarquables. Dans cette thèse, nous proposons deux nouveaux principes la vérification lexicale et la génération de cohorte afin d’attaquer les problèmes de la reconnaissance de l’écriture : i) le problème des grands lexiques et des décodages dirigés par le lexique ii) la problématique de combinaison de modèles optiques pour une meilleure reconnaissance iii) la nécessité de constituer de très grands ensembles de données étiquetées dans un contexte d’apprentissage profond. La vérification lexicale est une alternative aux décodages dirigés par le lexique peu étudiée à cause des faibles performances des modèles optiques historiques (HMM). Nous montrons dans cette thèse qu’elle constitue une alternative intéressante aux approches dirigées par le lexique lorsqu’elles s’appuient sur des modèles optiques très performants comme les RNN LSTM. La génération de cohorte permet de générer facilement et rapidement un grand nombre de réseaux récurrents complémentaires en un seul apprentissage. De ces deux techniques nous construisons et proposons un nouveau schéma de cascade pour la reconnaissance de mots isolés, une nouvelle combinaison au niveau ligne LV-ROVER et une nouvelle stratégie d’auto-apprentissage de RNN LSTM pour la reconnaissance de mots isolés. La cascade proposée permet de combiner avec la vérification lexicale des milliers de réseaux et atteint des résultats à l’état de l’art pour les bases Rimes et IAM. LV-ROVER a une complexité réduite par rapport à l’algorithme original ROVER et permet de combiner des centaines de réseaux sans modèle de langage tout en dépassant l’état de l’art pour la reconnaissance de lignes sur le jeu de donnéesRimes. Notre stratégie d’auto-apprentissage permet d’apprendre à partir d’un seul réseau BLSTM et sans paramètres grâce à la cohorte et la vérification lexicale, elle montre d’excellents résultats sur les bases Rimes et IAMState-of-the-art methods for handwriting recognition are based on LSTM recurrent neural networks (RNN) which achieve high performance recognition. In this thesis, we propose the lexicon verification and the cohort generation as two new building blocs to tackle the problem of handwriting recognition which are : i) the large vocabulary problem and the use of lexicon driven methods ii) the combination of multiple optical models iii) the need for large labeled dataset for training RNN. The lexicon verification is an alternative to the lexicon driven decoding process and can deal with lexicons of 3 millions words. The cohort generation is a method to get easily and quickly a large number of complementary recurrent neural networks extracted from a single training. From these two new techniques we build and propose a new cascade scheme for isolated word recognition, a new line level combination LV-ROVER and a new self-training strategy to train LSTM RNN for isolated handwritten words recognition. The proposed cascade combines thousands of LSTM RNN with lexicon verification and achieves state-of-the art word recognition performance on the Rimes and IAM datasets. The Lexicon Verified ROVER : LV-ROVER, has a reduce complexity compare to the original ROVER algorithm and combine hundreds of recognizers without language models while achieving state of the art for handwritten line text on the RIMES dataset. Our self-training strategy use both labeled and unlabeled data with the unlabeled data being self-labeled by its own lexicon verified predictions. The strategy enables self-training with a single BLSTM and show excellent results on the Rimes and Iam datasets
24
Dong, Yue. "Higher Order Neural Networks and Neural Networks for Stream Learning." Thesis, Université d'Ottawa / University of Ottawa, 2017. http://hdl.handle.net/10393/35731.
Full textAbstract:
The goal of this thesis is to explore some variations of neural networks. The thesis is mainly split into two parts: a variation of the shaping functions in neural networks and a variation of learning rules in neural networks.
In the first part, we mainly investigate polynomial perceptrons - a perceptron with a polynomial shaping function instead of a linear one. We prove the polynomial perceptron convergence theorem and illustrate the notion by showing that a higher order perceptron can learn the XOR function through empirical experiments with implementation. In the second part, we propose three models (SMLP, SA, SA2) for stream learning and anomaly detection in streams. The main technique allowing these models to perform at a level comparable to the state-of-the-art algorithms in stream learning is the learning rule used. We employ mini-batch gradient descent algorithm and stochastic gradient descent algorithm to speed up the models. In addition, the use of parallel processing with multi-threads makes the proposed methods highly efficient in dealing with streaming data. Our analysis shows that all models have linear runtime and constant memory requirement. We also demonstrate empirically that the proposed methods feature high detection rate, low false alarm rate, and fast response.
The paper on the first two models (SMLP, SA) is published in the 29th Canadian AI Conference and won the best paper award. The invited journal paper on the third model (SA2) for Computational Intelligence is under peer review.
25
Allen, T. J. "Optoelectronic neural networks." Thesis, University of Nottingham, 1996. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.362900.
Full text
26
Sloan, Cooper Stokes. "Neural bus networks." Thesis, Massachusetts Institute of Technology, 2018. http://hdl.handle.net/1721.1/119711.
Full textAbstract:
Thesis: M. Eng., Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, 2018.This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections.
Cataloged from student-submitted PDF version of thesis.
Includes bibliographical references (pages 65-68).
Bus schedules are unreliable, leaving passengers waiting and increasing commute times. This problem can be solved by modeling the traffic network, and delivering predicted arrival times to passengers. Research attempts to model traffic networks use historical, statistical and learning based models, with learning based models achieving the best results. This research compares several neural network architectures trained on historical data from Boston buses. Three models are trained: multilayer perceptron, convolutional neural network and recurrent neural network. Recurrent neural networks show the best performance when compared to feed forward models. This indicates that neural time series models are effective at modeling bus networks. The large amount of data available for training bus network models and the effectiveness of large neural networks at modeling this data show that great progress can be made in improving commutes for passengers.
by Cooper Stokes Sloan.
M. Eng.
27
Landry, Kenneth D. "Evolutionary neural networks." Thesis, Virginia Polytechnic Institute and State University, 1988. http://hdl.handle.net/10919/51904.
Full textAbstract:
To create neural networks that work, one needs to specify a structure and the interconnection weights between each pair of connected computing elements. The structure of a network can be selected by the designer depending on the application, although the selection of interconnection weights is a much larger problem. Algorithms have been developed to alter the weights slightly in order to produce the desired results. Learning algorithms such as Hebb's rule, the Delta rule and error propagation have been used, with success, to learn the appropriate weights. The major objection to this class of algorithms is that one cannot specify what is not desired in the network in addition to what is desired. An alternate method to learning the correct interconnection weights is to evolve a network in an environment that rewards "good” behavior and punishes "bad" behavior, This technique allows interesting networks to appear which otherwise may not be discovered by other methods of learning. In order to teach a network the correct weights, this approach simply needs a direction where an acceptable solution can be obtained rather than a complete answer to the problem.Master of Science
28
Boychenko, I. V., and G. I. Litvinenko. "Artificial neural networks." Thesis, Вид-во СумДУ, 2009. http://essuir.sumdu.edu.ua/handle/123456789/17044.
Full text
29
Chen, Prakoon. "The Neural Shell : a neural networks simulator." Connect to resource, 1989. http://rave.ohiolink.edu/etdc/view?acc%5Fnum=osu1228839518.
Full text
30
Bolt, George Ravuama. "Fault tolerance in artificial neural networks : are neural networks inherently fault tolerant?" Thesis, University of York, 1992. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.317683.
Full text
31
Flanagan, John Adrian. "Self-organising neural networks /." [S.l.] : [s.n.], 1994. http://library.epfl.ch/theses/?nr=1306.
Full text
32
Kocheisen, Michael. "Neural networks in photofinishing /." Zürich, 1997. http://e-collection.ethbib.ethz.ch/show?type=diss&nr=11985.
Full text
33
Wendemuth, Andreas. "Optimisation in neural networks." Thesis, University of Oxford, 1994. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.386749.
Full text
34
Corbett-Clark, Timothy Alexander. "Explanation from neural networks." Thesis, University of Oxford, 1998. http://ora.ox.ac.uk/objects/uuid:b94d702a-1243-4702-b751-68784c855ab2.
Full textAbstract:
Neural networks have frequently been found to give accurate solutions to hard classification problems. However neural networks do not make explained classifications because the class boundaries are implicitly defined by the network weights, and these weights do not lend themselves to simple analysis. Explanation is desirable because it gives problem insight both to the designer and to the user of the classifier. Many methods have been suggested for explaining the classification given by a neural network, but they all suffer from one or more of the following disadvantages: a lack of equivalence between the network and the explanation; the absence of a probability framework required to express the uncertainty present in the data; a restriction to problems with binary or coarsely discretised features; reliance on axis-aligned rules, which are intrinsically poor at describing the boundaries generated by neural networks. The structure of the solution presented in this thesis rests on the following steps: Train a standard neural network to estimate the class conditional probabilities. Bayes’ rule then defines the optimal class boundaries. Obtain an explicit representation of these class boundaries using a piece-wise linearisation technique. Note that the class boundaries are otherwise only implicitly defined by the network weights. Obtain a safe but possibly partial description of this explicit representation using rules based upon the city-block distance to a prototype pattern. The methods required to achieve the last two represent novel work which seeks to explain the answers given by a proven neural network solution to the classification problem.
35
Glackin, Cornelius. "Fuzzy spiking neural networks." Thesis, University of Ulster, 2009. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.505831.
Full text
36
Pritchett, William Christopher. "Neural networks for classification." Thesis, Monterey, California. Naval Postgraduate School, 1998. http://hdl.handle.net/10945/8735.
Full textAbstract:
CIVINSIn many applications, ranging from character recognition to signal detection to automatic target identification, the problem of signal classification is of interest. Often, for example, a signal is known to belong to one of a family of sets C sub 1..., C sub n and the goal is to classify the signal according to the set to which it belongs. The main purpose of this thesis is to show that under certain conditions placed on the sets, the theory of uniform approximation can be applied to solve this problem. Specifically, if we assume that sets C sub j are compact subsets of a normed linear space, several approaches using the Stone-Weierstrass theorem give us a specific structure for classification. This structure is a single hidden layer feedforward neural network. We then discuss the functions which comprise the elements of this neural network and give an example of an application
37
Menneer, Tamaryn Stable Ia. "Quantum artificial neural networks." Thesis, University of Exeter, 1999. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.286530.
Full text
38
Tattersall, Graham David. "Neural networks and generalisation." Thesis, University of East Anglia, 1998. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.266735.
Full text
39
Liu, Qian. "Deep spiking neural networks." Thesis, University of Manchester, 2018. https://www.research.manchester.ac.uk/portal/en/theses/deep-spiking-neural-networks(336e6a37-2a0b-41ff-9ffb-cca897220d6c).html.
Full textAbstract:
Neuromorphic Engineering (NE) has led to the development of biologically-inspired computer architectures whose long-term goal is to approach the performance of the human brain in terms of energy efficiency and cognitive capabilities. Although there are a number of neuromorphic platforms available for large-scale Spiking Neural Network (SNN) simulations, the problem of programming these brain-like machines to be competent in cognitive applications still remains unsolved. On the other hand, Deep Learning has emerged in Artificial Neural Network (ANN) research to dominate state-of-the-art solutions for cognitive tasks. Thus the main research problem emerges of understanding how to operate and train biologically-plausible SNNs to close the gap in cognitive capabilities between SNNs and ANNs. SNNs can be trained by first training an equivalent ANN and then transferring the tuned weights to the SNN. This method is called âoff-lineâ training, since it does not take place on an SNN directly, but rather on an ANN instead. However, previous work on such off-line training methods has struggled in terms of poor modelling accuracy of the spiking neurons and high computational complexity. In this thesis we propose a simple and novel activation function, Noisy Softplus (NSP), to closely model the response firing activity of biologically-plausible spiking neurons, and introduce a generalised off-line training method using the Parametric Activation Function (PAF) to map the abstract numerical values of the ANN to concrete physical units, such as current and firing rate in the SNN. Based on this generalised training method and its fine tuning, we achieve the state-of-the-art accuracy on the MNIST classification task using spiking neurons, 99.07%, on a deep spiking convolutional neural network (ConvNet). We then take a step forward to âon-lineâ training methods, where Deep Learning modules are trained purely on SNNs in an event-driven manner. Existing work has failed to provide SNNs with recognition accuracy equivalent to ANNs due to the lack of mathematical analysis. Thus we propose a formalised Spike-based Rate Multiplication (SRM) method which transforms the product of firing rates to the number of coincident spikes of a pair of rate-coded spike trains. Moreover, these coincident spikes can be captured by the Spike-Time-Dependent Plasticity (STDP) rule to update the weights between the neurons in an on-line, event-based, and biologically-plausible manner. Furthermore, we put forward solutions to reduce correlations between spike trains; thereby addressing the result of performance drop in on-line SNN training. The promising results of spiking Autoencoders (AEs) and Restricted Boltzmann Machines (SRBMs) exhibit equivalent, sometimes even superior, classification and reconstruction capabilities compared to their non-spiking counterparts. To provide meaningful comparisons between these proposed SNN models and other existing methods within this rapidly advancing field of NE, we propose a large dataset of spike-based visual stimuli and a corresponding evaluation methodology to estimate the overall performance of SNN models and their hardware implementations.
40
Kalchbrenner, Nal. "Encoder-decoder neural networks." Thesis, University of Oxford, 2017. http://ora.ox.ac.uk/objects/uuid:d56e48db-008b-4814-bd82-a5d612000de9.
Full textAbstract:
This thesis introduces the concept of an encoder-decoder neural network and develops architectures for the construction of such networks. Encoder-decoder neural networks are probabilistic conditional generative models of high-dimensional structured items such as natural language utterances and natural images. Encoder-decoder neural networks estimate a probability distribution over structured items belonging to a target set conditioned on structured items belonging to a source set. The distribution over structured items is factorized into a product of tractable conditional distributions over individual elements that compose the items. The networks estimate these conditional factors explicitly. We develop encoder-decoder neural networks for core tasks in natural language processing and natural image and video modelling. In Part I, we tackle the problem of sentence modelling and develop deep convolutional encoders to classify sentences; we extend these encoders to models of discourse. In Part II, we go beyond encoders to study the longstanding problem of translating from one human language to another. We lay the foundations of neural machine translation, a novel approach that views the entire translation process as a single encoder-decoder neural network. We propose a beam search procedure to search over the outputs of the decoder to produce a likely translation in the target language. Besides known recurrent decoders, we also propose a decoder architecture based solely on convolutional layers. Since the publication of these new foundations for machine translation in 2013, encoder-decoder translation models have been richly developed and have displaced traditional translation systems both in academic research and in large-scale industrial deployment. In services such as Google Translate these models process in the order of a billion translation queries a day. In Part III, we shift from the linguistic domain to the visual one to study distributions over natural images and videos. We describe two- and three- dimensional recurrent and convolutional decoder architectures and address the longstanding problem of learning a tractable distribution over high-dimensional natural images and videos, where the likely samples from the distribution are visually coherent. The empirical validation of encoder-decoder neural networks as state-of- the-art models of tasks ranging from machine translation to video prediction has a two-fold significance. On the one hand, it validates the notions of assigning probabilities to sentences or images and of learning a distribution over a natural language or a domain of natural images; it shows that a probabilistic principle of compositionality, whereby a high- dimensional item is composed from individual elements at the encoder side and whereby a corresponding item is decomposed into conditional factors over individual elements at the decoder side, is a general method for modelling cognition involving high-dimensional items; and it suggests that the relations between the elements are best learnt in an end-to-end fashion as non-linear functions in distributed space. On the other hand, the empirical success of the networks on the tasks characterizes the underlying cognitive processes themselves: a cognitive process as complex as translating from one language to another that takes a human a few seconds to perform correctly can be accurately modelled via a learnt non-linear deterministic function of distributed vectors in high-dimensional space.
41
Nyamapfene, Abel. "Unsupervised multimodal neural networks." Thesis, University of Surrey, 2006. http://epubs.surrey.ac.uk/844064/.
Full textAbstract:
We extend the in-situ Hebbian-linked SOMs network by Miikkulainen to come up with two unsupervised neural networks that learn the mapping between the individual modes of a multimodal dataset. The first network, the single-pass Hebbian linked SOMs network, extends the in-situ Hebbian-linked SOMs network by enabling the Hebbian link weights to be computed through one- shot learning. The second network, a modified counter propagation network, extends the unsupervised learning of crossmodal mappings by making it possible for only one self-organising map to implement the crossmodal mapping. The two proposed networks each have a smaller computation time and achieve lower crossmodal mean squared errors than the in-situ Hebbian- linked SOMs network when assessed on two bimodal datasets, an audio-acoustic speech utterance dataset and a phonological-semantics child utterance dataset. Of the three network architectures, the modified counterpropagation network achieves the highest percentage of correct classifications comparable to that of the LVQ-2 algorithm by Kohonen and the neural network for category learning by de Sa and Ballard in classification tasks using the audio-acoustic speech utterance dataset. To facilitate multimodal processing of temporal data, we propose a Temporal Hypermap neural network architecture that learns and recalls multiple temporal patterns in an unsupervised manner. The Temporal Hypermap introduces flexibility in the recall of temporal patterns - a stored temporal pattern can be retrieved by prompting the network with the temporal pattern's identity vector, whilst the incorporation of short term memory allows the recall of a temporal pattern, starting from the pattern item specified by contextual information up to the last item in the pattern sequence. Finally, we extend the connectionist modelling of child language acquisition in two important respects. First, we introduce the concept of multimodal representation of speech utterances at the one-word and two-word stage. This allows us to model child language at the one-word utterance stage with a single modified counterpropagation network, which is an improvement on previous models in which multiple networks are required to simulate the different aspects of speech at the one-word utterance stage. Secondly, we present, for the time, a connectionist model of the transition of child language from the one-word utterance stage to the two-word utterance stage. We achieve this using a gated multi-net comprising a modified counterpropagation network and a Temporal Hypermap.
42
Remmelzwaal, Leendert Amani. "Salience-affected neural networks." Master's thesis, University of Cape Town, 2009. http://hdl.handle.net/11427/12111.
Full textAbstract:
Includes abstract.Includes bibliographical references (leaves 46-49).
In this research, the salience of an entity refers to its state or quality of standing out, or receiving increased attention, relative to neighboring entities. By neighbouring entities we refer to both spatial (i.e. similar visual objects) and temporal (i.e. related concepts). In this research we model the effect of non-local connections using an ANN, creating a salience-affected neural network (SANN). We adapt an ANN to embody the capacity to respond to an input salience signal and to produce a reverse salience signal during testing. The input salience signal applied during training to each node has the effect of varying the node’s thresholds, depending on the activation level of the node. Each node produces a nodal reverse salience signal during testing (a measure of the threshold bias for the individual node). The reverse salience signal is defined as the summation of the nodal reverse salience signals observed at each node.
43
Cheung, Ka Kit. "Neural networks for optimization." HKBU Institutional Repository, 2001. http://repository.hkbu.edu.hk/etd_ra/291.
Full text
44
Thom, Markus [Verfasser]. "Sparse neural networks / Markus Thom." Ulm : Universität Ulm. Fakultät für Ingenieurwissenschaften und Informatik, 2015. http://d-nb.info/1067496319/34.
Full text
45
Bulot, Jean-Paul. "Echo cancellation via neural networks." Thesis, Georgia Institute of Technology, 2000. http://hdl.handle.net/1853/15407.
Full text
46
Frazão, Xavier Marques. "Deep learning model combination and regularization using convolutional neural networks." Master's thesis, 2014. http://hdl.handle.net/10400.6/5605.
Full textAbstract:
Convolutional neural networks (CNNs) were inspired by biology. They are hierarchical neural
networks whose convolutional layers alternate with subsampling layers, reminiscent of simple
and complex cells in the primary visual cortex [Fuk86a]. In the last years, CNNs have emerged
as a powerful machine learning model and achieved the best results in many object recognition
benchmarks [ZF13, HSK+12, LCY14, CMMS12].
In this dissertation, we introduce two new proposals for convolutional neural networks. The
first, is a method to combine the output probabilities of CNNs which we call Weighted
Convolutional Neural Network Ensemble. Each network has an associated weight that makes
networks with better performance have a greater influence at the time to classify a pattern
when compared to networks that performed worse. This new approach produces better results
than the common method that combines the networks doing just the average of the output
probabilities to make the predictions. The second, which we call DropAll, is a generalization
of two well-known methods for regularization of fully-connected layers within convolutional
neural networks, DropOut [HSK+12] and DropConnect [WZZ+13]. Applying these methods
amounts to sub-sampling a neural network by dropping units. When training with DropOut, a
randomly selected subset of the output layer’s activations are dropped, when training with
DropConnect we drop a randomly subsets of weights. With DropAll we can perform both methods
simultaneously. We show the validity of our proposals by improving the classification error on a
common image classification benchmark.
47
HUNG, YA-WEN, and 洪雅雯. "Short Term Wind Speed Forecasting by Combination of Convolutional Neural Networks and Bidirectional Long Short-Term Memory Neural Networks." Thesis, 2019. http://ndltd.ncl.edu.tw/handle/qc4qn3.
Full textAbstract:
碩士逢甲大學
電機工程學系
107
In recent years, as a result of Taiwan's energy transformation, it is expected that the share of renewable energy generation can be increased, and wind power generation is one of them. Taiwan is now actively promoting the establishment of offshore wind farms, but the impact of wind power generation, in addition to the wind turbine itself unit equipment, topography, weather factors, wind speed is also a major factor affecting wind power generation. With the progress and development of artificial intelligence, this paper uses deep learning to make predictions. Deep learning includes a variety of models, such as multi-layer perceptrons, convolutional neural networks, and recursive neural networks. Different neural networks are used to combine different models to analyze individual predictionresults and optimize their models, and select the best models among the trained models. The paper will input different weather data and use three methods to predict wind speed.The first method is convolutional neural network with long short-term memory, the second method is convolutional neural network with bidirectional long short-term memory and the third method is convolutional neural network with bidirectional long short-term memory and principal component analysis.Using the ability of convolutional neural network to extract images.Long short-term memory and bidirectional long short-term memory are time series neural networks, Using of the principal component analysis extracted weather data feature value modeling helps to reduce computation and storage capacity, and reduce the complexity of the model, thereby improving the widespread use of the model capacity.
48
ChengJiang-Yong and 承江永. "A study of Combination Wavelet Transform and Neural Networks for Islanding Detection." Thesis, 2006. http://ndltd.ncl.edu.tw/handle/03173006438619605975.
Full textAbstract:
碩士崑山科技大學
電機工程研究所
94
In this thesis, a wavelet transform combined with neural networks approach is proposed to detect the occurrence of islanding events in distributed generation systems. Due to the time-frequency localization capability exhibited by wavelet transform, the character of signals can be extracted. Therefore, the neural network can detect the islanding events. Besides this, the effectiveness of the method has been validated through different scenarios. Test results reveal the feasibility of the method for the application considered.
49
Hung, Tsai-Yuan, and 洪才元. "A TAIEX Forecasting Model with A Combination of Genetic Algorithms and Neural Networks." Thesis, 2008. http://ndltd.ncl.edu.tw/handle/53801006686845924215.
Full textAbstract:
碩士中原大學
資訊管理研究所
96
The forecast for the trend of Taiwan stock index should have an impact on the investors toward the relative market price. Based on the forecast of the pivot of market price, the investors can decide their plan about short-term market, futures or index options. However, there are several ways to forecast stock index. Neural networks have been proposed for modeling nonlinear data. In the artificial intelligence field, researchers frequently use neural network to forecast stock index, but the results usually cause a slightly price gap when the market encounters a huge stock shake. Therefore, this study uses a system simulation model to do the forecast the Taiwan stock index. We combine the genetic algorithms and the back-propagation neural network to construct artificial intelligence forecasting model. It also combines the experience rules to enhance the precision of the proposed model. As a result, the forecast of this models can increase the breadth of market trend as compared with the traditional neural network. It increases the accuracy rate, and reduces gaps. By using this model, investors can decide when and where to invest their money. Therefore, the model will be quite appealing if we can predict the market behaviour accurately. Furthermore, this prediction model can help individual investors to determine the correctness of expert’s knowledge and market research report in order to make a beneficial investment decision. Therefore, this forecasting model is important for investors to avoid investment risks and enlarge high-profit abilities on stock investment.
50
MU, Cing-Bo Tuan, and 端木慶博. "Implementation and Comparisons of Hyperbolic Tangent Function using Combination Circuit for Neural Networks." Thesis, 2015. http://ndltd.ncl.edu.tw/handle/27532135584260730037.
Full textAbstract:
碩士逢甲大學
資訊工程學系
103
Artificial Neural Networks (ANNs), which can also be referred Neural Networks (NNs). Artificial Neural Networks is a kind of parallel computing system which is an abstract simulation for basic features of the human brain, or any natural neural networks. In general, Artificial Neural Networks are implemented by software. Recently, the more research implementation using hardware for Artificial Neural Networks which cause hardware implementation is more efficient than software. The main blocks needed for hardware implementation of neural networks are adder, multiplier, and nonlinear activation function. Hyperbolic tangent and sigmoid are mostly using the nonlinear activation function. Both activation functions have an s-shaped curve while their output range is different. Currently there are several different methods to implement the activation function hardware, such as Piecewise Linear Approximation, Piecewise Non-Linear Approximation, Look up Tables, Bit-level Mapping Approximation. The bit-level mapping method approximates output based on a direct bit-level mapping of input. This method can be implemented using purely combinational circuits. In this paper, we implement the Hyperbolic Tangent Function using custom instructions in limit hardware still with good performance. In addition, we took more than one on Hyperbolic Tangent Function into a custom instruction. Then, we can choose according to our needs the custom instruction with different number of bits.