Dissertations / Theses on the topic 'Automatic networks'
To see the other types of publications on this topic, follow the link: Automatic 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 'Automatic 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
Grau, Leguia Marc. "Automatic reconstruction of complex dynamical networks." Doctoral thesis, Universitat Pompeu Fabra, 2019. http://hdl.handle.net/10803/666631.
Full textAbstract:
Un problema principal de la ciència de xarxes és com reconstruir (inferir) la topologia d’una xarxa real a partir de senyals mesurades de les seves unitats internes. Entendre la arquitectura d’una xarxa complexa és clau, no només per comprendre el seu funcionament, sinó també per predir i controlar el seu comportament. Els mètodes actualment disponibles es centren principalment en la detecció d’enllaços de xarxes no direccio- nals i sovint requereixen suposicions fortes sobre el sistema. Tanmateix, molts d’aquests mètodes no es poden aplicar a xarxes amb connexions direccionals. Per abordar aquest problema, en aquesta tesis ens centrarem en la inferència de xarxes direccionals. Concretament, desenvolupem un mètode de reconstrucció de xarxes basat en models que combina estadístiques de correlacions de derivades amb recuit simulat. A més, desenvolupem un mètode de reconstrucció basat en dades cimentat en una mesura d’interpedendència no lineal. Aquest mètode permet inferir la topologia de xarxes direccionals d’oscil.ladors caòtics de Lorenz per un subordre de la força d’acoblament i la densitat de la xarxa. Finalment, apliquem el mètode basat en dades a gravacions electroencefalogràfiques d’un pacient amb epilèpsia. Les xarxes cerebrals funcionals obtingu- des a partir d’aquest mètode són coherents amb la informació mèdica disponible.Un problema principal de la ciencia de redes es cómo reconstruir (inferir) la topología de una red real usando la señales medidas de sus unidades internas. Entender la arquitectura de redes complejas es clave, no solo para entender su funcionamiento pero también para predecir y controlar su comportamiento. Los métodos existentes se focalizan en la detección de redes no direccionales y normalmente requieren fuertes suposicio- nes sobre el sistema. Sin embargo, muchos de estos métodos no pueden ser aplicados en redes con conexiones direccionales. Para abordar este problema, en esta tesis estudiamos la reconstrucción de redes direccio- nales. En concreto, desarrollamos un método de reconstrucción basado en modelos que combina estadísticas de correlaciones de derivadas con recocido simulado. Además, desarrollamos un método basado en datos cimentado en una medida d’interdependencia no lineal. Este método permite inferir la topología de redes direccionales de osciladores caóticos de Lorenz para un subrango de la fuerza de acoplamiento y densidad de la red. Finalmente, aplicamos el método basado en datos a grabaciones electroencefalográficas de un paciente con epilepsia. Las redes cerebra- les funcionales obtenidas usando este método son consistentes con la información médica disponible.
A foremost problem in network science is how to reconstruct (infer) the topology of a real network from signals measured from its internal units. Grasping the architecture of complex networks is key, not only to understand their functioning, but also to predict and control their behaviour. Currently available methods largely focus on the detection of links of undirected networks and often require strong assumptions about the system. However, many of these methods cannot be applied to networks with directional connections. To address this problem, in this doctoral work we focus at the inference of directed networks. Specifically, we develop a model-based network reconstruction method that combines statistics of derivative-variable correlations with simulated annealing. We furthermore develop a data-driven reconstruction method based on a nonlinear interdependence measure. This method allows one to infer the topology of directed networks of chaotic Lorenz oscillators for a subrange of the coupling strength and link density. Finally, we apply the data-driven method to multichannel electroencephalographic recordings from an epilepsy patient. The functional brain networks obtained from this approach are consistent with the available medical information.
2
Glöckner, Reinhard Jörg. "Power transfer optimised automatic matching networks." Thesis, Northumbria University, 2009. http://nrl.northumbria.ac.uk/1622/.
Full textAbstract:
Matching networks are widely used to enhance active power transfer when radio frequency generators drive complex loads. The tuning of the network for varying loads typically involves searching for optimum matching conditions. However, improving the matching condition of the network does not necessarily indicate an increase in active power transfer. As an example, a 71 network with three adjustable elements may achieve comparable matches for a variety of elements' settings, each matching triple exhibiting a different transferred active power. Furthermore, the influence of the transmission lines used to connect the matching network to its source and load is rarely taken into account. The purpose of the work is to optimise the power gain of a narrowband matching sys-tem in the frequency range of 1.8 — 30 MHz. The system consists of a source, a match-ing network, a load and two interconnecting lines whose characteristic impedance is complex. The optimisation process involves optimum choice of the transmission lines' lengths and development of a matching strategy. Its objective is to ensure automatic and continuous adjustment of the matching network for optimum ac¬tive power transfer to its load while matching the network's input impedance to a resistive source. The network topologies employed are limited to the most common 71 and T networks consisting of two variable capacitors and one central inductor. Losses are assumed to be mainly caused by the inductor. An appropriate simple and synthetic model of the losses is proposed which is suitable for active power transfer optimisation. The model is validated against losses of inductors derived by different works. After choosing a proper network parametrisation and exact inclusion of the losses during network design, the losses of a network terminated by a resistance and de¬signed to match (exactly) a source resistance at its input are derived. Then its power gain is optimised by a proper choice of the network's parameter and the impact of changing the purely resistive termination to impedances exhibiting capacitive or in¬ductive imaginary parts is considered. An explicit solution is calculated for networks with a constant Q factor central inductor, its differences from the approximate solu¬tion (network elements designed as if the network would be lossless) are considered. Example diagrams are given illustrating those differences and power gain contour Smith charts are drawn for typical ranges of the L, iv, and T networks' elements. Combining the results of the different approaches yields an optimum matching strat¬egy. The losses of transmission lines connecting source and network of load and network are determined, where the lines' complex characteristic impedance is taken into account. Those losses are included in a power gain optimisation of the complete matching system. Finally, an experimental setup is designed under which the matching strategy of the network is tested and validated.
3
Li, Ping. "Neural networks for automatic arc welding." Thesis, University of Liverpool, 1995. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.284264.
Full text
4
Rivera, Corban G. "Automatic Reconstruction of the Building Blocks of Molecular Interaction Networks." Diss., Virginia Tech, 2008. http://hdl.handle.net/10919/28752.
Full textAbstract:
High-throughput whole-genome biological assays are highly intricate and difficult to interpret. The molecular interaction networks generated from evaluation of those experiments suggest that cellular functions are carried out by modules of interacting molecules. Reverse-engineering the modular structure of cellular interaction networks has the promise of significantly easing their analysis.
We hypothesize that:- cellular wiring diagrams can be decomposed into overlapping modules, where each module is a set of coherently-interacting molecules and
- a cell responds to a stress or a stimulus by appropriately modulating the activities of a subset of these modules.
- Given a wiring diagram and genome-wide gene expression data measured after the application of a stress or in a disease state, compute the active network of molecular interactions perturbed by the stress or the disease.
- Given the active networks for multiple stresses, stimuli, or diseases, compute a set of network legos, which are molecular modules with the property that each active network can be expressed as an appropriate combination of a subset of modules.
Ph. D.
5
Prager, Richard William. "Parallel processing networks for automatic speech recognition." Thesis, University of Cambridge, 1987. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.238443.
Full text
6
Lee, George J. (George Janbing) 1979. "Automatic service selection in dynamic wireless networks." Thesis, Massachusetts Institute of Technology, 2003. http://hdl.handle.net/1721.1/87434.
Full text
7
Crestel, Léopold. "Neural networks for automatic musical projective orchestration." Electronic Thesis or Diss., Sorbonne université, 2018. http://www.theses.fr/2018SORUS625.
Full textAbstract:
L’orchestration est l’art de composer un discours musical en combinant les timbres instrumentaux. La complexité de la discipline a longtemps été un frein à l’élaboration d’une théorie de l’orchestration. Ainsi, contrairement à l’harmonie ou au contrepoint qui s’appuient sur de solides constructions théoriques, l’orchestration reste de nos jours encore essentiellement enseignée à travers l’observation d’exemples canoniques. Notre objectif est de développer un système d’orchestration automatique de pièce pour piano en nous appuyant sur des méthodes d’apprentissage statistique. Nous nous focalisons sur le répertoire classique, cette technique d’écriture étant courante pour des compositeurs tels que Mozart ou Beethoven qui réalisaient d’abord une ébauche pianistique de leurs pièces orchestrales. En observant une large base de donnée de pièces pour orchestre et leurs réductions pour piano, nous évaluons l'aptitude des réseaux de neurones à apprendre les mécanismes complexes qui régissent l’orchestration. La vaste capacité d’apprentissage des architectures profondes semble adaptée à la difficulté du problème. Cependant, dans un contexte orchestrale, les représentations musicales symboliques traditionnelles donnent lieu à des vecteurs parcimonieux dans des espaces de grande dimension. Nous essayons donc de contourner ces difficultés en utilisant des méthodes auto-régressives et des fonctions d’erreur adaptées. Finalement, nous essayons de développer un système capable d'orchestrer en temps réel l'improvisation d'un pianisteOrchestration is the art of composing a musical discourse over a combinatorial set of instrumental possibilities. For centuries, musical orchestration has only been addressed in an empirical way, as a scientific theory of orchestration appears elusive. In this work, we attempt to build the first system for automatic projective orchestration, and to rely on machine learning. Hence, we start by formalizing this novel task. We focus our effort on projecting a piano piece onto a full symphonic orchestra, in the style of notable classic composers such as Mozart or Beethoven. Hence, the first objective is to design a system of live orchestration, which takes as input the sequence of chords played by a pianist and generate in real-time its orchestration. Afterwards, we relax the real-time constraints in order to use slower but more powerful models and to generate scores in a non-causal way, which is closer to the writing process of a human composer. By observing a large dataset of orchestral music written by composers and their reduction for piano, we hope to be able to capture through statistical learning methods the mechanisms involved in the orchestration of a piano piece. Deep neural networks seem to be a promising lead for their ability to model complex behaviour from a large dataset and in an unsupervised way. More specifically, in the challenging context of symbolic music which is characterized by a high-dimensional target space and few examples, we investigate autoregressive models. At the price of a slower generation process, auto-regressive models allow to account for more complex dependencies between the different elements of the score, which we believe to be of the foremost importance in the case of orchestration
8
Newman, Rhys A. "Automatic learning in computer vision." Thesis, University of Oxford, 1998. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.390526.
Full text
9
Winsnes, Casper. "Automatic Subcellular Protein Localization Using Deep Neural Networks." Thesis, KTH, Skolan för datavetenskap och kommunikation (CSC), 2016. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-189991.
Full textAbstract:
Protein localization is an important part in understanding the functionality of a protein. The current method of localizing proteins is to manually annotate microscopy images. This thesis investigates the feasibility of using deep artificial neural networks to automatically classify subcellular protein locations based on immunoflourescent images. We investigate the applicability in both single-label and multi-label classification, as well as cross cell line classification. We show that deep single-label neural networks can be used for protein localization with up to 73% accuracy. We also show the potential of deep multi-label neural networks for protein localization and cross cell line classification but conclude that more research is needed before we can say for certain that the method is applicable.
10
Howell, Andrew Jonathan. "Automatic face recognition using radial basis function networks." Thesis, University of Sussex, 1997. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.241635.
Full text
11
Pitkänen, P. (Perttu). "Automatic image quality enhancement using deep neural networks." Master's thesis, University of Oulu, 2019. http://jultika.oulu.fi/Record/nbnfioulu-201904101454.
Full textAbstract:
Abstract. Photo retouching can significantly improve image quality and it is considered an essential part of photography. Traditionally this task has been completed manually with special image enhancement software. However, recent research utilizing neural networks has been proven to perform better in the automated image enhancement task compared to traditional methods.
During the literature review of this thesis, multiple automatic neural-network-based image enhancement methods were studied, and one of these methods was chosen for closer examination and evaluation. The chosen network design has several appealing qualities such as the ability to learn both local and global enhancements, and its simple architecture constructed for efficient computational speed. This research proposes a novel dataset generation method for automated image enhancement research, and tests its usefulness with the chosen network design. This dataset generation method simulates commonly occurring photographic errors, and the original high-quality images can be used as the target data. This dataset design allows studying fixes for individual and combined aberrations. The underlying idea of this design choice is that the network would learn to fix these aberrations while producing aesthetically pleasing and consistent results.
The quantitative evaluation proved that the network can learn to counter these errors, and with greater effort, it could also learn to enhance all of these aspects simultaneously. Additionally, the network’s capability of learning local and portrait specific enhancement tasks were evaluated. The models can apply the effect successfully, but the results did not gain the same level of accuracy as with global enhancement tasks. According to the completed qualitative survey, the images enhanced by the proposed general enhancement model can successfully enhance the image quality, and it can perform better than some of the state-of-the-art image enhancement methods.Automaattinen kuvanlaadun parantaminen käyttämällä syviä neuroverkkoja. Tiivistelmä. Manuaalinen valokuvien käsittely voi parantaa kuvanlaatua huomattavasti ja sitä pidetään oleellisena osana valokuvausprosessia. Perinteisesti tätä tehtävää varten on käytetty erityisiä manuaalisesti operoitavia kuvankäsittelyohjelmia. Nykytutkimus on kuitenkin todistanut neuroverkkojen paremmuuden automaattisessa kuvanparannussovelluksissa perinteisiin menetelmiin verrattuna.
Tämän diplomityön kirjallisuuskatsauksessa tutkittiin useita neuroverkkopohjaisia kuvanparannusmenetelmiä, ja yksi näistä valittiin tarkempaa tutkimusta ja arviointia varten. Valitulla verkkomallilla on useita vetoavia ominaisuuksia, kuten paikallisten sekä globaalien kuvanparannusten oppiminen ja sen yksinkertaistettu arkkitehtuuri, joka on rakennettu tehokasta suoritusnopeutta varten. Tämä tutkimus esittää uuden opetusdatan generointimenetelmän automaattisia kuvanparannusmetodeja varten, ja testaa sen soveltuvuutta käyttämällä valittua neuroverkkorakennetta. Tämä opetusdatan generointimenetelmä simuloi usein esiintyviä valokuvauksellisia virheitä, ja alkuperäisiä korkealaatuisia kuvia voi käyttää opetuksen tavoitedatana. Tämän generointitavan avulla voitiin tutkia erillisten valokuvausvirheiden, sekä näiden yhdistelmän korjausta. Tämän menetelmän tarkoitus oli opettaa verkkoa korjaamaan erilaisia virheitä sekä tuottamaan esteettisesti miellyttäviä ja yhtenäisiä tuloksia.
Kvalitatiivinen arviointi todisti, että käytetty neuroverkko kykenee oppimaan erillisiä korjauksia näille virheille. Neuroverkko pystyy oppimaan myös mallin, joka korjaa kaikkia ennalta määrättyjä virheitä samanaikaisesti, mutta alhaisemmalla tarkkuudella. Lisäksi neuroverkon kyvykkyyttä oppia paikallisia muotokuvakohtaisia kuvanparannuksia arvioitiin. Koulutetut mallit pystyvät myös toteuttamaan paikallisen kuvanparannuksen onnistuneesti, mutta nämä mallit eivät yltäneet globaalien parannusten tasolle. Toteutetun kyselytutkimuksen mukaan esitetty yleisen kuvanparannuksen malli pystyy parantamaan kuvanlaatua onnistuneesti, sekä tuottaa parempia tuloksia kuin osa vertailluista kuvanparannustekniikoista.
12
Lin, Alvin. "Video Based Automatic Speech Recognition Using Neural Networks." DigitalCommons@CalPoly, 2020. https://digitalcommons.calpoly.edu/theses/2343.
Full textAbstract:
Neural network approaches have become popular in the field of automatic speech recognition (ASR). Most ASR methods use audio data to classify words. Lip reading ASR techniques utilize only video data, which compensates for noisy environments where audio may be compromised. A comprehensive approach, including the vetting of datasets and development of a preprocessing chain, to video-based ASR is developed. This approach will be based on neural networks, namely 3D convolutional neural networks (3D-CNN) and Long short-term memory (LSTM). These types of neural networks are designed to take in temporal data such as videos. Various combinations of different neural network architecture and preprocessing techniques are explored. The best performing neural network architecture, a CNN with bidirectional LSTM, compares favorably against recent works on video-based ASR.
13
Boley, Alexander. "Automatic wind turbine operation analysis through neural networks." Thesis, KTH, Skolan för elektro- och systemteknik (EES), 2017. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-214551.
Full textAbstract:
This master thesis handles the development of an automatic benchmarking program for wind turbines and the thesis works as the theoretical basis for this program. The program is created at the request of the power company OX2 who wanted this potential to be investigated. The mission given by the company is to: 1. to find a good key point indicator for the efficiency of a wind turbine, 2. to find an efficient way to assess this and 3. to write a program that does this automatically and continuously. The thesis determines with a study of previous research that the best method to utilize for these kinds of continuous analyses are artificial neural networks which can train themselves on historical data and then assess if the wind turbine is working better or worse than it should with regards to its history. This comparison between the neural network predicted operation and the actual operation works as the measurement of the efficiency, the key point indicator for how the turbine work compared to how it historically should operate. The program is based on this principle and is completely written in MATLAB. Further testing of the program found that the best variables to use are wind speed and the blade pitch angle as input variables for the neural network and active power as the target used as the variable to predict and assess the operation. The final program was able to be fully automated and integrated into the OX2 system thanks to the possibility to continuously import wind turbine data through APIs. In the final testing was the program able to identify 75% of the anomalies manually found in the half year and in the five turbines used for this thesis, the small anomalies not found manually but identified by the program excluded.Den här masteruppsatsen hanterar utvecklandet av ett automatiskt driftanalyseringsprogram för vindkraftverk och fungerar som det teoretiska underlaget för detta program. Programmet utvecklades på uppdrag av kraftbolaget OX2 som ville undersöka potentialen för ett sådant analysprogram i deras verksamhet. Uppdraget givet var att: 1. ta fram en bra indikator när det gäller den faktiska effektiviteten av ett vindkraftverk, 2. att hitta ett effektivt sätt att använda detta måttet i en analys där målet är att hitta avvikelser, och 3. skriva ett program som automatiskt kan använda måttet och metoden över tiden. Rapporten kommer via litteraturstudie fram till att tidigare forskning visar på att neurala nätverk är den mest lovande metoden för att genomföra sådan här analys. Dessa nätverk kan träna sig själva på historiska data och sedan analysera om vindturbinen arbetar bättre eller sämre än historiskt. Den här jämförelsen mellan den historiskt grundade förutspådda kraften ut och den faktiska kraften ut fungerar som kvalitetsmåttet på hur bra turbinen fungerar. Programmet är baserat på den här principen och är helt skriven i MATLAB. Vidare tester av programmet visar att de bästa variablerna att använda för att förutspå kraften ut är vindhastigheten och bladens vinkel mot vinden. Slutprogrammet var kapabelt att fullt automatiskt och integrerat i OX2s system identifiera 75% av alla avvikelser som manuellt hittats i ett halvårs data på de fem turbinerna använda för rapporten, småfel hittade av programmet men inte manuellt exkluderat.
14
Lenz, Lutz Henning. "Automatic Tuning of Integrated Filters Using Neural Networks." PDXScholar, 1993. https://pdxscholar.library.pdx.edu/open_access_etds/4604.
Full textAbstract:
Component values of integrated filters vary considerably due to· manufacturing tolerances and environmental changes. Thus it is of major importance that the components of an integrated filter be electronically tunable. The method explored in this thesis is the transconductance-C-method. A method of realizing higher-order filters is to use a cascade structure of second-order filters. In this context, a method of tuning second-order filters becomes important The research objective of this thesis is to determine if the Neural Network methodology can be used to facilitate the filter tuning process for a second-order filter (realized via the transconductance-C-method). Since this thesis is, at least to the knowledge of the author, the first effort in this direction, basic principles of filters and of Neural Networks [1-22] are presented. A control structure is proposed which comprises three parts: the filter, the Neural Network, and a digital spectrum analyzer. The digital spectrum analyzer sends a test signal to the filter and measures the magnitude of the output at 49 frequency samples. The Neural Network part includes a memory that stores the 49 sampled values of the nominal spectrum. ·A comparator subtracts the latter values from the measured (actual) values, and feeds them as input to the Neural Network. The outputs of the Neural Network are the values of the percentage tuning amount The adjusting device, which is envisioned as a component of the filter itself, translates the output of the Neural Network to adjustments in the value of the filter's transconductances. Experimental results provide a demonstration that the Neural Network methodology can be usefully applied to the above problem context. A feedforward, singlehidden layer Backpropagation Network reduces the manufacturing errors of up to 85% for the pole frequency and of up to 41% for the quality factor down to less than approximately 5% each. It is demonstrated that the method can be iterated to further reduce the error.
15
Gunnar, Anders. "Aspects of proactive traffic engineering in IP networks." Doctoral thesis, KTH, Reglerteknik, 2011. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-29558.
Full textAbstract:
To deliver a reliable communication service over the Internet it is essential for the network operator to manage the traffic situation in the network. The traffic situation is controlled by the routing function which determines what path traffic follows from source to destination. Current practices for setting routing parameters in IP networks are designed to be simple to manage. This can lead to congestion in parts of the network while other parts of the network are far from fully utilized. In this thesis we explore issues related to optimization of the routing function to balance load in the network and efficiently deliver a reliable communication service to the users. The optimization takes into account not only the traffic situation under normal operational conditions, but also traffic situations that appear under a wide variety of circumstances deviating from the nominal case. In order to balance load in the network knowledge of the traffic situations is needed. Consequently, in this thesis we investigate methods for efficient derivation of the traffic situation. The derivation is based on estimation of traffic demands from link load measurements. The advantage of using link load measurements is that they are easily obtained and consist of a limited amount of data that need to be processed. We evaluate and demonstrate how estimation based on link counts gives the operator a fast and accurate description of the traffic demands. For the evaluation we have access to a unique data set of complete traffic demands from an operational IP backbone. However, to honor service level agreements at all times the variability of the traffic needs to be accounted for in the load balancing. In addition, optimization techniques are often sensitive to errors and variations in input data. Hence, when an optimized routing setting is subjected to real traffic demands in the network, performance often deviate from what can be anticipated from the optimization. Thus, we identify and model different traffic uncertainties and describe how the routing setting can be optimized, not only for a nominal case, but for a wide range of different traffic situations that might appear in the network. Our results can be applied in MPLS enabled networks as well as in networks using link state routing protocols such as the widely used OSPF and IS-IS protocols. Only minor changes may be needed in current networks to implement our algorithms. The contributions of this thesis is that we: demonstrate that it is possible to estimate the traffic matrix with acceptable precision, and we develop methods and models for common traffic uncertainties to account for these uncertainties in the optimization of the routing configuration. In addition, we identify important properties in the structure of the traffic to successfully balance uncertain and varying traffic demands.QC 20110211
16
Amiri, Asgari J. N. "Automatic face recognition for television audience monitoring." Thesis, Brunel University, 1992. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.363285.
Full text
17
Braccini, Michele. "Automatic Design of Boolean Networks for Modelling Differentiation Trees." Master's thesis, Alma Mater Studiorum - Università di Bologna, 2016. http://amslaurea.unibo.it/10432/.
Full textAbstract:
Real living cell is a complex system governed by many process which are not yet fully understood: the process of cell differentiation is one of these.
In this thesis work we make use of a cell differentiation model to develop gene regulatory networks (Boolean networks) with desired differentiation dynamics.
To accomplish this task we have introduced techniques of automatic design and we have performed experiments using various differentiation trees.
The results obtained have shown that the developed algorithms, except the Random algorithm, are able to generate Boolean networks with interesting differentiation dynamics.
Moreover, we have presented some possible future applications and developments of the cell differentiation model in robotics and in medical research.
Understanding the mechanisms involved in biological cells can gives us the possibility to explain some not yet understood dangerous disease, i.e the cancer.
Le cellula è un sistema complesso governato da molti processi ancora non pienamente compresi: il differenziamento cellulare è uno di questi.
In questa tesi utilizziamo un modello di differenziamento cellulare per sviluppare reti di regolazione genica (reti Booleane) con dinamiche di differenziamento desiderate.
Per svolgere questo compito abbiamo introdotto tecniche di progettazione automatica e abbiamo eseguito esperimenti utilizzando vari alberi di differenziamento.
I risultati ottenuti hanno mostrato che gli algoritmi sviluppati, eccetto l'algoritmo Random, sono in grado di poter generare reti Booleane con dinamiche di differenziamento interessanti.
Inoltre, abbiamo presentato alcune possibili applicazioni e sviluppi futuri del modello di differenziamento in robotica e nella ricerca medica.
Capire i meccanismi alla base del funzionamento cellulare può fornirci la possibilità di spiegare patologie ancora oggi non comprese, come il cancro.
18
Barrera, Núñez Víctor Augusto. "Automatic diagnosis of voltage disturbances in power distribution networks." Doctoral thesis, Universitat de Girona, 2012. http://hdl.handle.net/10803/80944.
Full textAbstract:
This thesis proposes a framework for identifying the root-cause of a voltage disturbance, as well as, its source location (upstream/downstream) from the monitoring place. The framework works with three-phase voltage and current waveforms collected in radial distribution networks without distributed generation. Real-world and synthetic waveforms are used to test it. The framework involves features that are conceived based on electrical principles, and assuming some hypothesis on the analyzed phenomena. Features considered are based on waveforms and timestamp information. Multivariate analysis of variance and rule induction algorithms are applied to assess the amount of meaningful information explained by each feature, according to the root-cause of the disturbance and its source location. The obtained classification rates show that the proposed framework could be used for automatic diagnosis of voltage disturbances collected in radial distribution networks. Furthermore, the diagnostic results can be subsequently used for supporting power network operation, maintenance and planning.En esta tesis se propone una metodología para la identificación de la localización relativa (aguas arriba/abajo) y la causa de una perturbación eléctrica. La metodología utiliza las ondas trifásicas de tensión y de corriente registradas en redes de distribución radial sin presencia de generación distribuida. La metodología es validada utilizando perturbaciones eléctricas reales y simuladas. La metodología involucra atributos que han sido concebidos basándose en principios eléctricos e hipótesis de acuerdo a cada uno de los fenómenos eléctricos analizados. Se propusieron atributos tanto basados en la forma de onda como en la fecha de ocurrencia de la perturbación. La cantidad de información contenida y/o explicada por cada atributo es valorada mediante la aplicación del análisis multivariante de la varianza y algoritmos de extracción automática de reglas de decisión. Los resultados de clasificación muestran que la metodología propuesta puede ser utilizada para el diagnóstico automático de perturbaciones eléctricas registradas en redes de distribución radial. Los resultados de diagnóstico pueden ser utilizados para apoyar las tareas de operación, mantenimiento y planeamiento de las redes de distribución.
19
WOLF, ALEXANDRE STURMER. "AUTOMATIC ANALISYS OF ELECTROCARDIOGRAPHIC SIGNALS USING ARTIFICIAL NEURAL NETWORKS." PONTIFÍCIA UNIVERSIDADE CATÓLICA DO RIO DE JANEIRO, 2004. http://www.maxwell.vrac.puc-rio.br/Busca_etds.php?strSecao=resultado&nrSeq=4817@1.
Full textAbstract:
COORDENAÇÃO DE APERFEIÇOAMENTO DO PESSOAL DE ENSINO SUPERIORO objetivo dessa dissertação é o desenvolvimento de um algoritmo para a análise automática de sinais eletrocardiográficos, baseado em Redes Neurais Artificiais. O sistema é dividido em vários sub- programas utilizados para extrair informações do registro eletrocardiográfico de pacientes, informando a existência de anormalidades a partir da comparação dos valores obtidos com os valores de normalidade disponíveis na literatura biomédica. O programa utiliza 4 segundos do sinal de eletrocardiograma para uma análise classificatória inicial, verificando a viabilidade da extração de informações. Sendo possível esta extração, são obtidos os ciclos cardíacos existentes nesse sinal, e deles são extraídas informações quantitativas dos componentes de suas ondas, que posteriormente serão comparadas com faixas de normalidade por meio de um conjunto de regras heurísticas, indicando assim a possível presença de alterações morfológicas do registro. Esse programa pode ser utilizado em comunidades carentes para orientar a necessidade de encaminhamento a um especialista, cuja presença é rara na maior parte dos postos de atendimento generalista. Também pode auxiliar ao médico especialista, indicando de forma objetiva as possíveis alterações do registro eletrocardiográfico. Os resultados obtidos podem ser considerados satisfatórios, sendo que os valores são compatíveis com a sua natureza, principalmente no que diz respeito aos problemas de baixa razão sinal/ruído existente nos sinais analisados. Para verificação dos resultados de localização dos pontos inicial e final de cada componente do ECG, uma das métricas utilizadas foi o MAPE, obtendo-se, 19,44 por cento para onda P,4,85 por cento para o complexo QRS, 8,93 por cento para o início da onda T e 7,76 por cento para o final da onda T. Outra métrica utilizada para comparar os resultados obtidos com outro artigo, foi a Média Aritmética/Desvio Padrão, onde se obteve mi=-0,8264 ms e sigma=3,7037 ms para o início da onda P, mi=-1,5082 ms e sigma=2,2890 ms para o fim da onda P, mi=-0,2104 ms e sigma=3,2486 ms para o início do complexo QRS, mi=-0,4309 ms e sigma=3,9542 ms para o fim do complexo QRS, mi=-0,1926 ms e sigma=5,7413 ms para o início da onda T, mi=-0,3346 ms e sigma=6,3991 ms para o fim da onda T.
The objective of this dissertation is implementing an algorithm for automatic analysis of electrocardiographic signals, using Artificial Neural Networks. The system is divided into several subprograms that extract relevant information about the cardiac signal measured from patients, and points out possible abnormalities by comparison with normal values found in biomedical bibliography. The algorithm uses 4 seconds of the electrocardiogram signal for an initial classification, verifying the feasibility of information extraction. If the extraction is possible, the separate cardiac cycles are collected from the signal and quantitative values for the various components are determined. Finally, these values are compared with the normal values, indicating alterations of wave morphology. This algorithm has a clear relevance in low-income communities, being useful for an initial classification of the patients, being then forwarded to a cardiologist when ECG abnormalities are identified. Another potential use is in helping the cardiologist to automatically determine accurate values from the electrocardiographic register. The results can by considered satistactory, because the values are being compatible with their nature, mainly due to problems of low signal-to-noise ratio in analysed signals. For verification of the results, one metric used was the MAPE, obtaining 19,44 percent for the P wave, 4,85 percent for the QRS complex, 8,93 percent for the begining of the T wave and 7,76 percent for the end of T wave. Another metric used for comparing results with another article, was the Arithmetic Mean/Standard Deviation, obtaining u=-0,8264 ms and ó=3,7037 ms for the onset of the P wave, u=-1,5082 ms and ó=2,2890 ms for the offset of P wave, u=-0,2104 ms and ó=3,2486 ms for the onset of the QRS complex, u=-0,4309 ms and ó=3,9542 ms for the offset of the QRS complex, u=-0,1926 ms and ó=5,7413 ms for the onset of the T wave, u=-0,3346 ms and ó=6,3991 ms for the offset of the T wave.
20
Reikeras, Helge. "Audio-visual automatic speech recognition using Dynamic Bayesian Networks." Thesis, Stellenbosch : University of Stellenbosch, 2011. http://hdl.handle.net/10019.1/6777.
Full text
21
Lilja, Mathias. "Automatic Essay Scoring of Swedish Essays using Neural Networks." Thesis, Uppsala universitet, Statistiska institutionen, 2018. http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-352505.
Full textAbstract:
We propose a neural network-based system for automatically grading essays written in Swedish. Previous system either relies on laboriously crafted features extracted by human experts or are limited to essays written in English. By using different variations of Long Short-Term Memory (LSTM) networks, our system automatically learns the relation between Swedish high-school essays and their assigned score. Using all of the intermediate states from the LSTM network proved to be crucial in order to understand the essays. Furthermore, we evaluate different ways of representing words as dense vectors which ultimately have a substantial effect on the overall performance. We compare our results to the ones achieved by the first and previously only automatic essay scoring system designed for the Swedish language. Although no state-of-the-art performance is reached, indication of the potential from a neural based grading system is found.
22
Nordling, Love. "Evaluation of Generative Neural Networks for Automatic Defect Detection." Thesis, Uppsala universitet, Institutionen för informationsteknologi, 2020. http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-428411.
Full textAbstract:
Quality assurance of mass produced items is prone to errors when performedmanually by a human. This has created a need for an automated solution. Theemergence of deep neural networks has created systems that can be trained toclassify defect from non-defect items. However, to alleviate the need for largeamounts of manual labeling required for most classification networks, severalunsupervised methods have been used. This report evaluates the use of a deepautoencoder for unsupervised defect detection. Furthermore is the use of anautoencoder compared to applying inpainting and a generate adversarialnetwork(GAN) for the same task.The report finds that the autoencoder used could find the largest of defects testedbut not the smaller ones. It is also shown that neither use of inpainting nor a GANimproved on the autoencoder result. It is of note however that it was a naiveimplementation of inpainting and the GAN and they were lacking some state of theart aspects.
23
Chen, Hsinchun, and K. J. Lynch. "Automatic Construction of Networks of Concepts Characterizing Document Databases." IEEE, 1992. http://hdl.handle.net/10150/105175.
Full textAbstract:
Artificial Intelligence Lab, Department of MIS, University of ArizonaThe results of a study that involved the creation of knowledge bases of concepts from large, operational textual databases are reported. Two East-bloc computing knowledge bases, both based on a semantic network structure, were created automatically using two statistical algorithms. With the help of four East-bloc computing experts, we evaluated the two knowledge bases in detail in a concept-association experiment based on recall and recognition tests. In the experiment, one of the knowledge bases that exhibited the asymmetric link property out-performed all four experts in recalling relevant concepts in East-bloc computing. The knowledge base, which contained about 20,O00 concepts (nodes) and 280,O00 weighted relationships (links), was incorporated as a thesaurus-like component into an intelligent retrieval system. The system allowed users to perform semantics-based information management and information retrieval via interactive, conceptual relevance feedback.
24
Nguyễn, Hoàng Hà. "Automatic reconstruction of realistic road networks from GIS data." Thesis, Aix-Marseille, 2016. http://www.theses.fr/2016AIXM4007/document.
Full textAbstract:
La reconstruction de routes est un sujet important dans le domaine de la modélisation 3D. Nous proposons dans cette thèse des méthodes pour construire des modèles de réseaux routiers réalistes à partir de données SIG.Tout d'abord, les problèmes de la reconstruction d'un simple axe de route à partir d'une polyligne sont étudiés. Nous présentons un modèle de courbe par morceaux G1 qui est non seulement fidèle à l'axe de la route réelle mais aussi pratique et pas onéreux à obtenir. Notre algorithme Least Square Growing crée, comme dans le génie civil, une courbe horizontale et une courbe verticale, puis les combine pour produire un axe routier 3D approchant la polyligne. Traiter individuellement des polylignes conduit à des discontinuités aux intersections des routes. C’est pourquoi nous introduisons une procédure pour détecter les relations entre les routes, afin de proposer un processus global pour reconstruire tous les axes routiers avec prenant en compte les nouvelles contraintes sur les extrémités des routes.Enfin, sur la base de l'axe de la route et les propriétés résultant de la route dans la base de données SIG, nous définissons un modèle mathématique de la surface de la route en respectant les contraintes essentielles de surfaces routières réelles. Pour produire une géométrie représentant la surface de la route finale, nous construisons un maillage grossier de la carte d'élévation du terrain d'entrée, que nous subdivisons adaptativement le long de l'axe de la route, puis ajustons l'altitude des sommets concernés à la valeur définie par le modèle mathématique de la route afin de parvenir à une correspondance correcte entre le terrain et la vraie routeRoad reconstruction is an important topic in 3D modeling. Recently, the steady development of many critical-accurate applications has posed a high demand for realistic road models, taking into account road-design constraints selected from civil engineering. We propose in this dissertation methods for building realistic road network models from GIS data.Firstly, problems of single road axis reconstruction from a polyline are addressed. We present a novel G1-piecewise-curve model which is not only faithful to the real road axis but also convenient and cheap to render. Our Least Square Growing Algorithm creates, as in civil engineering, an horizontal and a vertical curves, then combines them to produce a 3D road axis fitting well the polyline. Processing individual polyline will leads to the discontinuities at road intersections so we introduce a procedure to detect road relations, then we propose a global process to reconstruct all road axes with the considerations on further constraints of road ends.Finally, based upon the resulting road axis and road properties in the GIS database, we define a mathematical road surface model respecting the essential constraints of real road surfaces. To produce a geometry representing the final road surface, we build a coarse mesh from the input terrain highmap, subdivide it adaptively along the road axis, then adjust the altitude of concerning vertices to the value defined by the mathematical model in order to attain a correct mapping between the terrain and the real road
25
Brückner, Jörg. "Automatic pattern recognition and learning for information systems." Thesis, University of Sussex, 1995. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.262632.
Full text
26
Fairgrieve, Andrew. "The application of neural networks in active suspension." Thesis, Loughborough University, 2003. https://dspace.lboro.ac.uk/2134/34234.
Full textAbstract:
This thesis considers the application of neural networks to automotive suspension systems. In particular their ability to learn non-linear feedback control relationships. The speed of processing, once trained, means that neural networks open up new opportunities and allow increased complexity in the control strategies employed. The suitability of neural networks for this task is demonstrated here using multilayer perceptron, (MLP) feed forward neural networks applied to a quarter vehicle simulation model. Initially neural networks are trained from a training data set created using a non-linear optimal control strategy, the complexity of which prohibits its direct use. They are shown to be successful in learning the relationship between the current system states and the optimal control.
27
Silos, Sánchez Àngel. "Automatic fault location in electrical distribution networks with distributed generation." Doctoral thesis, Universitat Politècnica de Catalunya, 2018. http://hdl.handle.net/10803/650824.
Full textAbstract:
Nowadays the electrical network is continuously evolving due to the increasing deployment of Information Technologies and the Distribution Energy Resources. This scenario affects directly to the quality of service in the electrical distribution networks. For this reason, the Power Quality is a key important concern to make the electrical network evolve towards a Smart Grid.
Power quality is defined through three important focal points: availability, wave quality and commercial quality. The presence of the Distribution Energy Resources in the current electrical distribution network is showing a new scenario where the fault detection is more complex due to the flow current is in both directions.
This thesis is focused in the analysis of several methods to locate a fault in electrical distribution network and also how the current communication standards can improve considerably this fault location. It is important to remark that the main contribution of this thesis is in the analysis of several propositions and algorithms to enhance the fault location in a distribution network using the current Intelligent Electronic Device with international standards such as IEC 61850. All of these algorithms have been focused to work in a mesh distribution networks.
Another important contribution of this thesis is in the adaptive protection system in order to isolate correctly the fault in a ring system distribution. Although this proposition could be extended to a mesh network where the elements of the network can operate under a fault.
Finally, the thesis concludes that the use of communication standards and Internet of Things with current developed Intelligent Electronic Devices technology can contribute significantly to enhance the current and future electrical network distribution.La xarxa elèctrica evoluciona contínuament a causa del creixent desplegament de les Tecnologies de la Informació i dels Recursos Energètics Distribuïts. Aquest escenari afecta directament a la qualitat de servei de les xarxes de distribució elèctrica. Per aquest motiu, el mantenir i millorar el nivell de qualitat d'energia és un punt clau per fer evolucionar la xarxa elèctrica cap a una xarxa Smart Grid. Aquesta qualitat de l'energia es defineix per medi de de tres punts importants: disponibilitat, qualitat d'ona i qualitat comercial. La presència dels Recursos Energètics Distribuïts mostra un nou escenari en què la detecció de defectes es complica afectant a la disponibilitat del servei. Aquesta tesi es centra principalment en l'anàlisi de diversos mètodes per localitzar un defecte a la xarxa de distribució elèctrica i també en com l'ús dels estàndards de comunicació actuals poden contribuir considerablement a la localització del defecte. És important remarcar que la principal contribució d'aquest document ha estat en l'anàlisi de diverses proposicions i algoritmes per millorar la localització de faltes en una xarxa de distribució utilitzant Dispositius Electrònics Intel·ligents amb estàndards internacionals com l'IEC 61850. Tots aquests algoritmes han estat definits per treballar en xarxes de distribució mallades. Una altra contribució important d'aquesta tesi es troba en el sistema de protecció adaptatiu per tal d'aïllar correctament el defecte en una distribució del sistema d'anell amb interruptors automàtics. Aquesta proposta es podria ampliar a una xarxa mallada. Finalment, la tesi conclou amb que l'ús d'estàndards de comunicació i l'Internet of Things en combinació amb Dispositius Electrònics Intel·ligents, desenvolupats actualment, poden contribuir significativament a millorar la distribució de la xarxa elèctrica actual i futura.
28
Martínez, Marrodán Diego. "Automatic Scheduling for schools : Scalability of the Hopfield Neural Networks." Thesis, KTH, Skolan för datavetenskap och kommunikation (CSC), 2015. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-166589.
Full textAbstract:
This project is focused on evaluating, in terms of real time needed to find a solution, the scalability of Hopfield Neural Networks, a Machine Learning method, applied to a common problem that every educational institution has to deal with at least once in every academic year, timetabling. With this purpose, the problem is first introduced. And secondly, in the background, the concept of "constraint" is presented, to continue with a brief explanation of Artificial Neural Networks, the state of the art and more specifically, how Hopfield Neural Networks are characterized. The formulation, modifications used, and the algorithm are presented. This algorithm will be implemented in MATLAB, and it will be run on data sets of different sizes. The results obtained for the presented data sets are presented in a table and graphs, to later discuss these results. In this discussion, it is found that the time spent to get a solution could scale quadratically with respect to the size of the problem, but there is not statistical evidence to this hypothesis. Finally, the conclusion is that Hopfield Neural Networks could have a good scalability if the hypothesis worked for bigger data sets, and some future work in the field is presented, like using sparse matrices for the implementation of the problem, or studying the scalability of Hopfield Neural Networks in other kinds of scheduling.
29
Tuerk, Christine M. "Automatic speech synthesis using auditory transforms and artificial neural networks." Thesis, University of Cambridge, 1992. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.385362.
Full text
30
Dielmann, Alfred. "Automatic recognition of multiparty human interactions using dynamic Bayesian networks." Thesis, University of Edinburgh, 2009. http://hdl.handle.net/1842/4022.
Full textAbstract:
Relating statistical machine learning approaches to the automatic analysis of multiparty communicative events, such as meetings, is an ambitious research area. We have investigated automatic meeting segmentation both in terms of “Meeting Actions” and “Dialogue Acts”. Dialogue acts model the discourse structure at a fine grained level highlighting individual speaker intentions. Group meeting actions describe the same process at a coarse level, highlighting interactions between different meeting participants and showing overall group intentions. A framework based on probabilistic graphical models such as dynamic Bayesian networks (DBNs) has been investigated for both tasks. Our first set of experiments is concerned with the segmentation and structuring of meetings (recorded using multiple cameras and microphones) into sequences of group meeting actions such as monologue, discussion and presentation. We outline four families of multimodal features based on speaker turns, lexical transcription, prosody, and visual motion that are extracted from the raw audio and video recordings. We relate these lowlevel multimodal features to complex group behaviours proposing a multistreammodelling framework based on dynamic Bayesian networks. Later experiments are concerned with the automatic recognition of Dialogue Acts (DAs) in multiparty conversational speech. We present a joint generative approach based on a switching DBN for DA recognition in which segmentation and classification of DAs are carried out in parallel. This approach models a set of features, related to lexical content and prosody, and incorporates a weighted interpolated factored language model. In conjunction with this joint generative model, we have also investigated the use of a discriminative approach, based on conditional random fields, to perform a reclassification of the segmented DAs. The DBN based approach yielded significant improvements when applied both to the meeting action and the dialogue act recognition task. On both tasks, the DBN framework provided an effective factorisation of the state-space and a flexible infrastructure able to integrate a heterogeneous set of resources such as continuous and discrete multimodal features, and statistical language models. Although our experiments have been principally targeted on multiparty meetings; features, models, and methodologies developed in this thesis can be employed for a wide range of applications. Moreover both group meeting actions and DAs offer valuable insights about the current conversational context providing valuable cues and features for several related research areas such as speaker addressing and focus of attention modelling, automatic speech recognition and understanding, topic and decision detection.
31
Navaneethan, Senthivadivelu. "Automatic fault location system for low voltage underground distribution networks." Thesis, University of Strathclyde, 2003. http://oleg.lib.strath.ac.uk:80/R/?func=dbin-jump-full&object_id=21540.
Full textAbstract:
This thesis presents a novel approach to automating Time Domain Reflectrometry (TDR) waveform acquisition and automatic TDR based fault location in Low Voltage (450-1000V) Underground Distribution Networks (LVUDNs). First, the types of faults that occur in LVUDN and previously available fault location techniques are discussed and their relative advantages and limitations described. Adaptive Filter theory, Wavelet Transform Theory and Fuzzy Logic are presented. Software is developed to automate: checking of the test lead connections, adjusting the internal balance network to match the cable surge impedance, blown fuse detection and backfeed identification, auto recording and storage of data, and voltage and current triggering for transient faults. Software is also developed for both direct and remote control of the instrument via a standard telephone line, GSM modem or direct serial link. Adaptive and fuzzy based, and wavelet based automatic fault location systems are developed. Both systems pre-process the TDR waveforms by using a simple thresholding technique to identify single phase tees and to locate three phase faults. The adaptive and fuzzy based system uses an adaptive filter to produce a composite waveform from the healthy and faulty TDR waveforms and the fault distance is calculated using the composite waveform. If the result produces more than one possible fault distance either from the TDR waveforms or the error waveforms, the system uses fuzzy reasoning to find a common fault distance. In the wavelet based fault location process the TDR waveforms are split into four multi-scales before applying the adaptive filtering and calculating the fault distance using a selected scale. To improve the accuracy of fault distance calculation, local mean and gradient techniques are used in the adaptive and fuzzy based fault location system and latter technique is used in the wavelet enhanced fault location system. The performances of both systems were tested using data from a cable model and from real LVUDNs and gave an accuracy of ±4.3m of the actual fault distance.
32
Khare, Vineet R. "Automatic problem decomposition using co-evolution and modular neural networks." Thesis, University of Birmingham, 2006. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.435412.
Full text
33
Rosten, David Paul 1967. "Automatic design of a decision tree classifier employing neural networks." Thesis, The University of Arizona, 1991. http://hdl.handle.net/10150/277881.
Full textAbstract:
Pattern recognition problems involve two main issues: feature formulation and classifier design. This thesis is concerned with the latter. Numerous algorithms for the design of pattern recognition systems have been published, and the algorithm detailed herein is a new approach--specific to the design of decision tree classifiers. It involves a top-down strategy, optimizing the root node decision and then subsequently its children. To assess various pattern space partitions, the Tie statistical distance measure quantified the separability of potential cluster groupings. Additionally, a separate neural network was employed at each of the tree decision nodes. Results from the application of this methodology to the regional labeling of panchromatic images suggest it is a suitable approach.
34
Garces, Freddy. "Dynamic neural networks for approximate input- output linearisation-decoupling of dynamic systems." Thesis, University of Reading, 2000. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.368662.
Full text
35
Tucker, Allan Brice James. "The automatic explanation of Multivariate Time Series with large time lags." Thesis, Birkbeck (University of London), 2001. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.246924.
Full text
36
Reynolds, James H. "Spoken letter recognition with neural networks." Thesis, University of Oxford, 1991. http://ora.ox.ac.uk/objects/uuid:b30872a7-7bd8-437f-bd3a-649de981d352.
Full textAbstract:
Neural networks have recently been applied to real-world speech recognition problems with a great deal of success. This thesis developes a strategy for optimising a neural network known as the Radial Basis Function classifier (RBF), on a large spoken letter recognition problem designed by British Telecom Research Laboratories. The strategy developed can be viewed as a compromise between a fully adaptive approach involving prohibitively large amounts of computation, and a heuristic approach resulting in poor generalisation. A value for the optimal number of kernel functions is suggested, and methods for determining the positions of the centres and the values of the width parameters are provided. During the evolution of the optimisation strategy it was demonstrated that spatial organisation of the centres does not adversely affect the ability of the classifier to generalise. An RBF employing the optimisation strategy achieved a lower error rate than a multilayer perceptron and two traditional static pattern classifiers on the same problem. The error rate of the RBF was very close to the theoretical minimum error rate obtainable with an optimal Bayes classifier. In addition to error rate, the performance of the classifiers was assessed in terms of the computational requirements of training and classification, illustrating the significant trade-off between computational investment in training and level of generalisation achieved. The error rate of the RBF was compared with that of a well established method of dynamic classification to examine whether non-linear time normalisation of word patterns was advantageous to generalisation. It was demonstrated that the dynamic classifier was better suited to small-scale speech recognition problems, and the RBF to speaker-independent speech recognition problems. The dynamic classifier was then combined with a neural network algorithm, greatly reducing its computational requirement without significantly increasing its error rate. This system was then extended into a novel system for visual feedback therapy in which speech is visualised as a moving trajectory on a computer screen.
37
Zhang, Dian. "Transceiver-free object tracking in wireless sensor networks /." View abstract or full-text, 2006. http://library.ust.hk/cgi/db/thesis.pl?COMP%202006%20ZHANG.
Full text
38
Fahmy, Maged Mohamed Mahoud. "Application of computer vision to automatic vehicle identification." Thesis, University of Newcastle Upon Tyne, 1994. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.244940.
Full text
39
Zhong, Xin. "Speech coding and transmission for improved automatic recognition in communication networks." Diss., Available online, Georgia Institute of Technology, 2004:, 2003. http://etd.gatech.edu/theses/available/etd-04072004-180252/unrestricted/zhong%5Fxin%5F200312%5Fphd.pdf.
Full text
40
Rosa, Junior Leomar Soares da. "Automatic generation and evaluation of transistor networks in different logic styles." reponame:Biblioteca Digital de Teses e Dissertações da UFRGS, 2008. http://hdl.handle.net/10183/61869.
Full textAbstract:
O projeto e o desenvolvimento de circuitos integrados é um dos mais importantes e aquecidos segmentos da indústria eletrônica da atualidade. Neste cenário, ferramentas de automação têm possibilitado aos projetistas manipular uma elevada quantidade de transistores em circuitos cada vez mais complexos, diminuindo, assim, o tempo de projeto. Em especial, ferramentas de síntese lógica têm contribuído significativamente para reduzir o ciclo de desenvolvimento. Na metodologia de projeto full-custom, cada bloco funcional tem sua geração realizada de forma manual, desde a implementação das redes de transistores até a geração do leiaute. Entretanto, esta tarefa é extremamente custosa em tempo de projeto. Neste contexto, torna-se confortável ter a disposição algoritmos dedicados para derivar redes de transistores automaticamente. Diversos tipos de arranjos de transistores são encontrados na literatura. Estas diferentes redes de transistores apresentam diferentes comportamentos em termos de consumo de área, consumo de potência e velocidade. Desta forma, não apenas a geração automática de redes de transistores é importante, mas também técnicas automatizadas para avaliar e comparar estas distintas redes de chaves é de fundamental importância para guiar o projetista que deseja alcançar implementações de circuitos eficientes. Estas avaliações não precisam ser necessariamente processos custosos de caracterização elétrica. Elas podem ser realizadas através de estimativas capazes de fornecer informações acuradas sobre o comportamento das redes. Esta idéia pode ser utilizada por projetistas que desejam gerar e avaliar potenciais soluções em redes de transistores para alimentar fluxos standard-cell (utilizando bibliotecas de células), ou por aqueles que utilizam a abordagem de mapeamento tecnológico library-free (fazendo uso de geradores de células). Neste contexto, este trabalho apresenta um gerador automático de redes de transistores capaz de fornecer diferentes tipos de redes em diversos estilos lógicos. Para comparar as redes geradas, algumas técnicas de estimativa são empregadas. Comparações são realizadas sobre conjuntos distintos de funções Booleanas, demonstrando as vantagens da utilização de lógicas alternativas em relação ao difundido padrão CMOS.Currently, VLSI design has established a dominant role in the electronics industry. Automated tools have enabled designers to manipulate more transistors on a design project and shorten the design cycle. In particular, logic synthesis tools have contributed significantly to reduce the design cycle time. In full-custom designs, manual generation of transistor netlists for each functional block is performed, but this is an extremely time-consuming task. In this sense, it becomes comfortable to have efficient algorithms to derive transistor networks automatically. There are several kinds of transistor networks arrangements. These different networks present different behaviors in terms of area, delay and power consumption. Thus, not only automatic transistor networks generation is important, but also an automated technique to evaluate and to compare the distinct switch networks is fundamental to guide designers that need to achieve efficient circuit implementations. This evaluation not necessarily needs to be an expensive electrical characterization process. It can be obtained through estimation processes capable of delivering good information about the logic cells behavior. This idea is useful for those designers that desire to generate and to evaluate potential transistor network implementations to feed standard-cell flow designs (using cell libraries), or for those designers who target the use of library-free technology mapping concept (using automatic cells generators). In this context, this work presents an automated transistor network generator able to delivery different kinds of networks in several logic styles. In order to compare the obtained networks, some estimation techniques are employed. A comparison is done over a set of Boolean function benchmarks, showing the advantages of using alternative logic styles over the traditional Complementary Series-Parallel CMOS (CSP CMOS).
41
Sleep, Jonathan. "Automatic Music Transcription with Convolutional Neural Networks using Intuitive Filter Shapes." DigitalCommons@CalPoly, 2017. https://digitalcommons.calpoly.edu/theses/1803.
Full textAbstract:
This thesis explores the challenge of automatic music transcription with a combination of digital signal processing and machine learning methods. Automatic music transcription is important for musicians who can't do it themselves or find it tedious. We start with an existing model, designed by Sigtia, Benetos and Dixon, and develop it in a number of original ways. We find that by using convolutional neural networks with filter shapes more tailored for spectrogram data, we see better and faster transcription results when evaluating the new model on a dataset of classical piano music. We also find that employing better practices shows improved results. Finally, we open-source our test bed for pre-processing, training, and testing the models to assist in future research.
42
Lavenius, Axel. "Automatic identification of northern pike (Exos Lucius) with convolutional neural networks." Thesis, Uppsala universitet, Institutionen för geovetenskaper, 2020. http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-418639.
Full textAbstract:
The population of northern pike in the Baltic sea has seen a drasticdecrease in numbers in the last couple of decades. The reasons for this are believed to be many, but the majority of them are most likely anthropogenic. Today, many measures are being taken to prevent further decline of pike populations, ranging from nutrient runoff control to habitat restoration. This inevitably gives rise to the problem addressed in this project, namely: how can we best monitor pike populations so that it is possible to accurately assess and verify the effects of these measures over the coming decades? Pike is currently monitored in Sweden by employing expensive and ineffective manual methods of individual marking of pike by a handful of experts. This project provides evidence that such methods could be replaced by a Convolutional Neural Network (CNN), an automatic artificial intelligence system, which can be taught how to identify pike individuals based on their unique patterns. A neural net simulates the functions of neurons in the human brain, which allows it to perform a range of tasks, while a CNN is a neural net specialized for this type of visual recognition task. The results show that the CNN trained in this project can identify pike individuals in the provided data set with upwards of 90% accuracy, with much potential for improvement.
43
Noblía, Matilda. "Automatic Anomaly Detection in Graphical User Interfaces Using Deep Neural Networks." Thesis, KTH, Skolan för elektroteknik och datavetenskap (EECS), 2019. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-264457.
Full textAbstract:
The automatic detection of code errors is a ubiquitous part of the quality assurance process performed during software development. However, graphical errors that may occur in user interfaces are often detected manually. This report examines if deep neural networks (DNNs), may be used to automatically detect two common types of anomalies present in a graphical user interface. The results point towards this being the case for the particular dataset used in this report.Automatisk detektering av kodfel är standard i kvalitetsarbetet som utförs vid mjukvaruveckling. Grafiska fel som kan uppstå i användargränssnitt upptäcks dock ofta manuellt. Den här rapporten undersöker ifall djupa neurala nätverk kan användas för att automatiskt detektera två vanliga fel som uppstår i användargränssnitt. Resultaten indikerar att så är fallet åtminstone för det specifika dataset som används.
44
Neggaz, Mohammed Yessin. "Automatic classification of dynamic graphs." Thesis, Bordeaux, 2016. http://www.theses.fr/2016BORD0169/document.
Full textAbstract:
Les réseaux dynamiques sont constitués d’entités établissant des contacts les unes avec les autres dans le temps. Un défi majeur dans les réseaux dynamiques est de prédire les modèles de mobilité et de décider si l’évolution de la topologie satisfait aux exigences du succès d’un algorithme donné. Les types de dynamique résultant de ces réseaux sont variés en échelle et en nature. Par exemple,certains de ces réseaux restent connexes tout le temps; d’autres sont toujours déconnectés mais offrent toujours une sorte de connexité dans le temps et dans l’espace(connexité temporelle); d’autres sont connexes de manière récurrente, périodique,etc. Tous ces contextes peuvent être représentés sous forme de classes de graphes dynamiques correspondant à des conditions nécessaires et/ou suffisantes pour des problèmes ou algorithmes distribués donnés. Étant donné un graphe dynamique,une question naturelle est de savoir à quelles classes appartient ce graphe. Dans ce travail, nous apportons une contribution à l’automatisation de la classification de graphes dynamiques. Nous proposons des stratégies pour tester l’appartenance d’un graphe dynamique à une classe donnée et nous définissons un cadre générique pour le test de propriétés dans les graphes dynamiques. Nous explorons également le cas où aucune propriété sur le graphe n’est garantie, à travers l’étude du problème de maintien d’une forêt d’arbres couvrants dans un graphe dynamiqueDynamic networks consist of entities making contact over time with one another. A major challenge in dynamic networks is to predict mobility patterns and decide whether the evolution of the topology satisfies requirements for the successof a given algorithm. The types of dynamics resulting from these networks are varied in scale and nature. For instance, some of these networks remain connected at all times; others are always disconnected but still offer some kind of connectivity over time and space (temporal connectivity); others are recurrently connected,periodic, etc. All of these contexts can be represented as dynamic graph classes corresponding to necessary or sufficient conditions for given distributed problems or algorithms. Given a dynamic graph, a natural question to ask is to which of the classes this graph belongs. In this work we provide a contribution to the automation of dynamic graphs classification. We provide strategies for testing membership of a dynamic graph to a given class and a generic framework to test properties in dynamic graphs. We also attempt to understand what can still be done in a context where no property on the graph is guaranteed through the distributed problem of maintaining a spanning forest in highly dynamic graphs
45
Ramani, Vipin. "Reconfigurable control using polynomial neural networks." Diss., Georgia Institute of Technology, 1996. http://hdl.handle.net/1853/13297.
Full text
46
Byrne, Carlton B. "Assembly task identification and strategy development using expert systems and neural networks." Thesis, Cardiff University, 1993. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.266650.
Full text
47
Dias, De Macedo Filho Antonio. "Microwave neural networks and fuzzy classifiers for ES systems." Thesis, University College London (University of London), 1996. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.244066.
Full text
48
Wan, Yan. "New paradigms for design and control of dynamical networks." Pullman, Wash. : Washington State University, 2009. http://www.dissertations.wsu.edu/Dissertations/Spring2009/y_wan_022509.pdf.
Full textAbstract:
Thesis (Ph. D.)--Washington State University, May 2009.Title from PDF title page (viewed on Apr. 2, 2009). "School of Electrical Engineering & Computer Science." Includes bibliographical references (p. 433-454).
49
Pan, Junfeng. "Learning-based localization in wireless and sensor networks /." View abstract or full-text, 2007. http://library.ust.hk/cgi/db/thesis.pl?CSED%202007%20PAN.
Full text
50
Casini, Luca. "Automatic Music Generation Using Variational Autoencoders." Master's thesis, Alma Mater Studiorum - Università di Bologna, 2018. http://amslaurea.unibo.it/16137/.
Full textAbstract:
The aim of the thesis is the design and evaluation of a generative model based on deep learning for creating symbolic music. Music, and art in general, pose interesting problems from a machine learning standpoint as they have structure and coherence both locally and globally and also have semantic content that goes beyond the mere structural problems.
Working on challenges like those can give insight on other problems in the machine learning world. Historically algorithmic music generation focused on structure and was achieved through the use of Markov models or by defining, often manually, a set of strict rules to be followed. In recent years the availability of large amounts of data and cheap computational power led to the resurgence of Artificial Neural Networks (ANN). Deep Learning is machine learning based on ANN with many stacked layers and is improving state of the art in many fields, including generative models.
This thesis focuses on Variational Autoencoders(VAE), a type of neural network where the input is mapped to a lower-dimensional code that is fit to a Gaussian distribution and then tries to reconstruct it minimizing the error. The distribution can be easily sampled allowing to generate and interpolate data in the latent space. Autoencoders can use any type of network to encode and decode the input, we will use Convolutional Neural Network (CNN) and Recurrent Neural Netowrks (RNN).
Since the quality of music and art in general is deeply subjective and what seems pleasing to one may not be for another we will try to determine the “best” model by conducting a survey and asking the participants to rate their enjoyment of music and whether or not they think each sample to be composed by a human or AI.