Spelling suggestions: "subject:"programming""
31 |
Isometry and convexity in dimensionality reductionVasiloglou, Nikolaos 30 March 2009 (has links)
The size of data generated every year follows an exponential growth. The number of data points as well as the dimensions have increased dramatically the past 15 years. The gap between the demand from the industry in data processing and the solutions provided by the machine learning community is increasing. Despite the growth in memory and computational power, advanced statistical processing on the order of gigabytes is beyond any possibility. Most sophisticated Machine Learning algorithms require at least quadratic complexity. With the current computer model architecture, algorithms with higher complexity than linear O(N) or O(N logN) are not considered practical. Dimensionality reduction is a challenging problem in machine learning. Often data represented as multidimensional points happen to have high dimensionality. It turns out that the information they carry can be expressed with much less dimensions. Moreover the reduced dimensions of the data can have better interpretability than the original ones. There is a great variety of dimensionality reduction algorithms under the theory of Manifold Learning. Most of the methods such as Isomap, Local Linear Embedding, Local Tangent Space Alignment, Diffusion Maps etc. have been extensively studied under the framework of Kernel Principal Component Analysis (KPCA). In this dissertation we study two current state of the art dimensionality reduction methods, Maximum Variance Unfolding (MVU) and Non-Negative Matrix Factorization (NMF). These two dimensionality reduction methods do not fit under the umbrella of Kernel PCA. MVU is cast as a Semidefinite Program, a modern convex nonlinear optimization algorithm, that offers more flexibility and power compared to iv KPCA. Although MVU and NMF seem to be two disconnected problems, we show that there is a connection between them. Both are special cases of a general nonlinear factorization algorithm that we developed. Two aspects of the algorithms are of particular interest: computational complexity and interpretability. In other words computational complexity answers the question of how fast we can find the best solution of MVU/NMF for large data volumes. Since we are dealing with optimization programs, we need to find the global optimum. Global optimum is strongly connected with the convexity of the problem. Interpretability is strongly connected with local isometry1 that gives meaning in relationships between data points. Another aspect of interpretability is association of data with labeled information. The contributions of this thesis are the following:
1. MVU is modified so that it can scale more efficient. Results are shown on 1 million speech datasets. Limitations of the method are highlighted.
2. An algorithm for fast computations for the furthest neighbors is presented for the first time in the literature.
3. Construction of optimal kernels for Kernel Density Estimation with modern convex programming is presented. For the first time we show that the Leave One Cross Validation (LOOCV) function is quasi-concave.
4. For the first time NMF is formulated as a convex optimization problem
5. An algorithm for the problem of Completely Positive Matrix Factorization is presented.
6. A hybrid algorithm of MVU and NMF the isoNMF is presented combining advantages of both methods.
7. The Isometric Separation Maps (ISM) a variation of MVU that contains classification information is presented.
8. Large scale nonlinear dimensional analysis on the TIMIT speech database is performed.
9. A general nonlinear factorization algorithm is presented based on sequential convex programming. Despite the efforts to scale the proposed methods up to 1 million data points in reasonable time, the gap between the industrial demand and the current state of the art is still orders of magnitude wide.
|
32 |
Contratos formais para derivação e verificação de componentes paralelos / Formal contracts for derivation and verification of parallel componentesMarcilon, Thiago Braga January 2012 (has links)
MARCILON, Thiago Braga. Contratos formais para derivação e verificação de componentes paralelos. 2012. 156 f. Dissertação (Mestrado em ciência da computação)- Universidade Federal do Ceará, Fortaleza-CE, 2012. / Submitted by Elineudson Ribeiro (elineudsonr@gmail.com) on 2016-07-12T18:10:13Z
No. of bitstreams: 1
2012_dis_tbmarcilon.pdf: 1474623 bytes, checksum: e61fb07d3a835088ff5a53d4eb684f2d (MD5) / Approved for entry into archive by Rocilda Sales (rocilda@ufc.br) on 2016-07-22T12:39:34Z (GMT) No. of bitstreams: 1
2012_dis_tbmarcilon.pdf: 1474623 bytes, checksum: e61fb07d3a835088ff5a53d4eb684f2d (MD5) / Made available in DSpace on 2016-07-22T12:39:34Z (GMT). No. of bitstreams: 1
2012_dis_tbmarcilon.pdf: 1474623 bytes, checksum: e61fb07d3a835088ff5a53d4eb684f2d (MD5)
Previous issue date: 2012 / The use of cloud computing to offer High Performance Computing (HPC) services has been widely discussed in the academia and industry. In this respect, this dissertation is included in the context of designing a cloud computing platform for the development of component-based parallel computing applications, referred as cloud of components. Many important challenges about using the cloud of components relate to parallel programming, an error-prone task due to synchronization issues, which may lead to abortion and production of incorrect data during execution of applications, and the inefficient use of computational resources. These problems may be very relevant in the case of long running applications with tight timelines to obtain critical results, quite common in the context of HPC. One possible solution to these problems is the formal analysis of the behavior of the components of an application through the cloud services, before their execution. Thus, the users of the components may know if a component can be safely used in their application. In this scenario, formal methods becomes useful. In this dissertation, it is proposed a process for specification and derivation of parallel components implementation for the cloud of components. This process involves the formal specification of the components behavior through contracts described using the Circus formal specification language. Then, through a refinement and translation process, which takes the contract as a start point, one may produce an implementation of a component that may execute on a parallel computing platform. Through this process, it becomes possible to offer guarantees to developers about the components behavior in their applications. To validate the proposed idea, the process is applied to contracts that have been described based on two benchmarks belonging to the NAS Parallel Benchmarks, widely adopted in HPC for evaluate the performance of parallel programming and computing platforms. / A aplicação de nuvens computacionais para oferecer serviços de Computação de Alto Desempenho (CAD) é um assunto bastante discutido no meio acadêmico e industrial. Esta dissertação está inserida no contexto do projeto de uma nuvem computacional para o desenvolvimento e execução de aplicações de CAD baseadas em componentes paralelos, doravante denominada nuvem de componentes. Um dos principais desafios na sua utilização consiste no suporte à programação paralela, tarefa bastante suscetível à erros, pois tais erros podem levar, ao longo do desenvolvimento, a problemas de sincronização de processos, que podem causar abortamento da execução e a produção de dados incorretos, bem como a problemas relacionados ao uso ineficiente dos recursos computacionais. É importante que tais problemas sejam tratados no caso de aplicações de longa duração cujo respeito a um cronograma para obtenção de resultados é crítico, aplicações estas bastante comuns no contexto de CAD. Uma possível solução para tais problemas consiste na verificação do comportamento e das propriedades dos componentes na nuvem, antes que seja feita a sua execução, tornando possível que os usuários dos componentes da nuvem saibam se um componente pode ser utilizado com segurança em sua aplicação. Nesse cenário, o uso de métodos formais surge como uma alternativa atraente. A contribuição desta dissertação consiste em um processo de derivação e verificação de propriedades de componentes na nuvem. Tal processo envolve a especificação formal do comportamento dos componentes por meio de contratos descritos pela linguagem Circus. Então, através de um processo de refinamento e tradução tendo como ponto de partida o contrato, chega-se à implementação de um componente para execução sobre uma plataforma de computação paralela. Através desse processo, torna-se possível oferecer garantias aos desenvolvedores em relação ao comportamento dos componentes no contexto de suas aplicações. Para a prova de conceito, o processo é aplicado sobre a especificação "papel-e-caneta" de dois benchmarks do NAS Parallel Benchmarks, IS e CG, bastante difundidos na área de CAD.
|
33 |
Výuka algoritmizace na základní škole - aktuální stav / Current situation in teaching of algorithmization in secondary schoolBROMOVÁ, Jana January 2012 (has links)
The aim of this thesis is mapping of current state of teaching of algorithmization in secondary school. Basic information about the subject is given in theoretical part - a relation between teaching of algorithmization and RVP ZV, and basic assupmtions of teaching of algorithmization: a historical aspect, methods, aims and possibilities of teaching of algorithmization (and an approach on teaching). The essential part of teaching of algorithmization is an education support - a textbook about algorithmization and programming, programming support, projects and competitions for students. Practical part is a pedagogical research. This research consists of five themes A. A quantitative research of teaching of algorithmization in secondary school in České Budějovice. B. A review of extra curricural teaching of algorithmization in ČB. C. A statewide mapping of teaching of algorithmization in Czech Republic. D. A statewide research on a success rate of the competition ?Bobřík informatiky? (informatics beaver). E. A teacher's experience with a textbook.
|
34 |
Contratos formais para derivaÃÃo e verificaÃÃo de componentes paralelos. / Formal Contracts for Derivation and Verification of Parallel ComponentesThiago Braga Marcilon 20 September 2012 (has links)
Conselho Nacional de Desenvolvimento CientÃfico e TecnolÃgico / A aplicaÃÃo de nuvens computacionais para oferecer serviÃos de ComputaÃÃo de Alto Desempenho (CAD) à um assunto bastante discutido no meio acadÃmico e industrial. Esta dissertaÃÃo està inserida no contexto do projeto de uma nuvem computacional para o desenvolvimento e execuÃÃo de aplicaÃÃes de CAD baseadas em componentes paralelos, doravante denominada nuvem de componentes. Um dos principais desafios na sua utilizaÃÃo consiste no suporte à programaÃÃo paralela, tarefa bastante suscetÃvel à erros, pois tais erros podem levar, ao longo do desenvolvimento, a problemas de sincronizaÃÃo de processos, que podem causar abortamento da execuÃÃo e a produÃÃo de dados incorretos, bem como a problemas relacionados ao uso ineficiente dos recursos computacionais. à importante que tais problemas sejam tratados no caso de aplicaÃÃes de longa duraÃÃo cujo respeito a um cronograma para obtenÃÃo de resultados à crÃtico, aplicaÃÃes estas bastante comuns no contexto de CAD. Uma possÃvel soluÃÃo para tais problemas consiste na verificaÃÃo do comportamento e das propriedades dos componentes na nuvem, antes que seja feita a sua execuÃÃo, tornando possÃvel que os usuÃrios dos componentes da nuvem saibam se um componente pode ser utilizado com seguranÃa em sua aplicaÃÃo. Nesse cenÃrio, o uso de mÃtodos formais surge como uma alternativa atraente. A contribuiÃÃo desta dissertaÃÃo consiste em um processo de derivaÃÃo e verificaÃÃo de propriedades de componentes na nuvem. Tal processo envolve a especificaÃÃo formal do comportamento dos componentes por meio de contratos descritos pela linguagem Circus. EntÃo, atravÃs de um processo de refinamento e traduÃÃo tendo como ponto de partida o contrato, chega-se à implementaÃÃo de um componente para execuÃÃo sobre uma plataforma de computaÃÃo paralela. AtravÃs desse processo, torna-se possÃvel oferecer garantias aos desenvolvedores em relaÃÃo ao comportamento dos componentes no contexto de suas aplicaÃÃes. Para a prova de conceito, o processo à aplicado sobre a especificaÃÃo "papel-e-caneta" de dois benchmarks do NAS Parallel Benchmarks, IS e CG, bastante difundidos na Ãrea de CAD. / The use of cloud computing to offer High Performance Computing (HPC) services has been widely discussed in the academia and industry. In this respect, this dissertation is included in the context of designing a cloud computing platform for the development of component-based parallel computing applications, referred as cloud of components. Many important challenges about using the cloud of components relate to parallel programming, an error-prone task due to synchronization issues, which may lead to abortion and production of incorrect data during execution of applications, and the inefficient use of computational resources. These problems may be very relevant in the case of long running applications with tight timelines to obtain critical results, quite common in the context of HPC. One possible solution to these problems is the formal analysis of the behavior of the components of an application through the cloud services, before their execution. Thus, the users of the components may know if a component can be safely used in their application. In this scenario, formal methods becomes useful. In this dissertation, it is proposed a process for specification and derivation of parallel components implementation for the cloud of components. This process involves the formal specification of the components behavior through contracts described using the Circus formal specification language. Then, through a refinement and translation process, which takes the contract as a start point, one may produce an implementation of a component that may execute on a parallel computing platform. Through this process, it becomes possible to offer guarantees to developers about the components behavior in their applications. To validate the proposed idea, the process is applied to contracts that have been described based on two benchmarks belonging to the NAS Parallel Benchmarks, widely adopted in HPC for evaluate the performance of parallel programming and computing platforms.
|
35 |
Métodos de otimização para a resolução do problema do despacho hidrotérmico-eólico em função da não linearidade da altura de quedaSORIANO, Guilherme Delgado 19 July 2016 (has links)
Submitted by Irene Nascimento (irene.kessia@ufpe.br) on 2017-01-27T17:45:09Z
No. of bitstreams: 2
license_rdf: 1232 bytes, checksum: 66e71c371cc565284e70f40736c94386 (MD5)
Versão Digital SORIANO M279.pdf: 2376439 bytes, checksum: b48b7385113bece4aa5c13da969e9642 (MD5) / Made available in DSpace on 2017-01-27T17:45:09Z (GMT). No. of bitstreams: 2
license_rdf: 1232 bytes, checksum: 66e71c371cc565284e70f40736c94386 (MD5)
Versão Digital SORIANO M279.pdf: 2376439 bytes, checksum: b48b7385113bece4aa5c13da969e9642 (MD5)
Previous issue date: 2016-07-19 / FACEPE / O Sistema Interligado Nacional (SIN), singular pelo seu nível de contribuição de fontes
renováveis e caracterizado como hidrotérmico, é regido de acordo com o planejamento da
operação. Levando em conta custo nulo para a geração de origem hidráulica, e custo elevado
devido ao alto preço dos combustíveis empregados para a geração térmica, o objetivo
econômico do planejamento da operação é substituir sempre que possível a geração de origem
térmica por geração de origem hidráulica. Devido à complexidade para resolução do despacho
hidrotérmico, é apresentado nessa dissertação o software dhoVisual com um modelo de
otimização não linear, baseado na altura de queda das hidrelétricas com reservatório, que toma
a decisão mais apropriada para a operação do SIN em cada período considerado. Para isso, foi
adotado o método de Pontos Interiores comparando casos otimizados de forma não linear com
casos otimizados linearmente. / The National Interconnected System (SIN), unique for its level of contribution from renewable
sources and characterized as hydrothermal, is regulated in accordance with the planning of the
operation. Taking into account no cost for the generation of hydroelectric and high cost due to
the high price of fuel used for heat generation, the economic objective of the planning of the
operation is always to replace, as possible, the generation of thermal generation by generation
of hydraulic origin. Concerning the complexity to solve the hydrothermal dispatch, it will be
presented in this dissertation the dhoVisual software with a nonlinear optimization model
based on the head variation of hydropower with reservoir, taking the most appropriate decision
for SIN operation in each period considered. For this, it was adopted the Interior Point method
comparing optimized cases of non-linearity with cases optimized linearly.
|
36 |
Experiments with Vehicle PlatooningWoldu, Essayas Gebrewahid, Jokhio, Fareed Ahmed January 2010 (has links)
This thesis is concerned with an experimental platform for studying cooperative driving and techniques for embedded systems programming. Cooperative driving systems use vehicle-to-vehicle and vehicle-to-infrastructure communication for safe, smooth and efficient transportation. Cooperative driving systems are considered as a promising solution for traffic situations such as blind crossings. For the thesis work we use a robotic vehicle known as PIE (Platform for Intelligent Embedded Systems) equipped with a wireless communication device, electrical motors and controlled via a SAM7-P256 development board. For the infrastructure side we use a SAM7-P256 development board equipped with nRF24l01. Vehicle to vehicle and base station to vehicle communication is established and different platooning scenarios are implemented. The scenarios are similar to platooning scenarios from the Grand Cooperative Driving Challenge GCDC1. The performance of the platoon control algorithm is measured in terms of throughput (a measure of string stability), smoothness and safety, where the safety requirements serve as pass/fail criteria.
|
37 |
Adaptation of Legacy Codes to Context-Aware Composition using Aspect-Oriented Programming for Data Representation ConversionSotsenko, Alisa January 2013 (has links)
Different computation problem domains such as sorting, matrix multiplication, etc. usually require different data representations and algorithms variants implementations in order to be adapted and re-designed to context-aware composition (CAC). Context-aware composition is a technique for the design of applications that can adapt its behavior according to changes in the program. We considered two application domains: matrix multiplication and graph algorithms (DFS algorithm in particular). The main problem in the implementation of the representation mechanisms applied in these problem domains is time spent on the data representation conversion that in the end should not influence the application performance. This thesis work presents a flexible aspect-based architecture that includes the data structure representation adaptation in order to reduce implementation efforts required for adaptation different application domains. Although, manual approach has small overhead 4-10% for different problems compared to the AOP-based approach, experiments show that the manual adaptation to CAC requires on average three times more programming effort in terms of lines of code than AOP-based approach. Moreover, the AOP-based approach showed the average speed-up over baseline algorithms that use standard data structures of 2.1.
|
38 |
Métodos de otimização para resolução do problema do despacho hidrotérmico considerando a Geração Eólica em três patamares de cargaMELO, Rodrigo Nunes de 17 June 2016 (has links)
Submitted by Fabio Sobreira Campos da Costa (fabio.sobreira@ufpe.br) on 2017-07-11T12:50:27Z
No. of bitstreams: 2
license_rdf: 811 bytes, checksum: e39d27027a6cc9cb039ad269a5db8e34 (MD5)
DissertaçãoRodrigoNunesdeMelo.pdf: 5589879 bytes, checksum: e851e203ef8b11b9711c04c1c34a7ede (MD5) / Made available in DSpace on 2017-07-11T12:50:27Z (GMT). No. of bitstreams: 2
license_rdf: 811 bytes, checksum: e39d27027a6cc9cb039ad269a5db8e34 (MD5)
DissertaçãoRodrigoNunesdeMelo.pdf: 5589879 bytes, checksum: e851e203ef8b11b9711c04c1c34a7ede (MD5)
Previous issue date: 2016-06-17 / FACEPE / O planejamento da operação de um sistema elétrico de potência está diretamente relacionado
com o despacho de usinas hidrelétricas e termelétricas. As características geográficas do Brasil
contribuem para que o parque gerador seja predominantemente hidráulico. Devido à grande
dimensão dos sistemas elétricos, a otimização do problema de despacho hidrotérmico é uma
tarefa extremamente complexa que pode ser realizada de modo eficiente, buscando otimizar a
operação dos reservatório das usinas hidráulicas, onde o objetivo é a redução do custo na
geração térmica necessária para atendimento à carga e eventuais déficits de energia, além de
maior nível de segurança. O presente trabalho aborda o desenvolvimento e a implementação
de um software para resolução do problema do despacho hidrotérmico em três patamares de
carga a ser atendido. Neste trabalho o problema de despacho foi formulado como um
problema de programação linear, que por sua vez foi solucionado pelos métodos de pontos
interiores primal-dual e preditor-corretor de barreira logarítmica. O trabalho faz uma avaliação
do desempenho computacional dos métodos implementados e do método LINPROG presente
no software Matlab® na solução do problema de planejamento da operação em larga escala,
para horizontes de cinco e de dez anos. As simulações foram feitas baseados em dados do
Plano Decenal de Energia (PDE) 2022 e apresentaram desempenhos satisfatórios. / The operational planning of electric power systems is directly related to the dispatch of
hydroelectric and thermal power plants. The Brazilian electric energy park is a predominantly
hydraulic system, due to its geographic characteristics. Due to the large size of the electrical
systems, the optimization of the hydrothermal dispatch problem is an extremely complex task
that can be carried out efficiently, seeking to optimize the operation of the reservoir in the
hydroelectric plants aimed at reducing the cost of the necessary thermal generation to meet the
load and possible energy deficits, and a high level of security. This work discusses the
development and implementation of a software to solve the hydrothermal dispatch problem in
three load steps. In this dissertation the hydrothermal dispatching problem is formulated as a
linear programming program, which in term is solved by the following methods of interior
point: primal-dual and predictor-corrector with logarithmic barrier. This work provides an
evaluation of the computational performance of the implemented methods and LINPROG,
presents in the software Matlab®, to solve a large scale operational planning problem, for
horizons of five and ten years. The simulation were made based on data from the “Plano
Decenal de Energia (PDE) 2022” and showed satisfactory performance.
|
39 |
Balanceamento de linhas de produção com trabalhadores deficientes e máquinas paralelas / Balancing assembly lines with disabled workers and parallel workstationsFelipe Francisco Bezerra Araújo 03 May 2011 (has links)
O problema de balanceamento de linhas de produção e designação de trabalhadores (ALWABP, do inglês Assembly Line Worker Assignment and Balancing Problem ) é uma extensão do problema simples de balanceamento de linhas na qual os tempos de execução de tarefas são dependentes dos trabalhadores. Este problema tem sua motivação prática oriunda de linhas de produção com trabalhadores deficientes. Neste trabalho, focamos em uma extensão para o problema que permite lidar com a possibilidade de layouts de linhas com estações em paralelo. Além disso, estudamos uma segunda variante do problema que permite que os trabalhadores colaborem uns com os outros em uma mesma estação. Apresentamos formulações matemáticas para ambas as versões do problema. Adicionalmente, a partir de adaptações de métodos existentes para o ALWABP serial, desenvolvemos heurísticas construtivas para o ALWABP paralelo e ALWABP colaborativo. Testes computacionais em instâncias da literatura e uma análise detalhada dos resultados são apresentados / The assembly line and worker assignment balancing problem is an extension of the classical simple assembly line balancing problem where task execution times are workerdependent. This problem is motivated by the situation faced when balancing assembly lines with disabled workers. In this study, we focus on an extension for the problem with the goal of including the possibility of having layouts with parallel workstations. We also study a second variation for this problem that allows workers collaborating with each other. We present mathematical formulations for both problems. Moreover, constructive heuristics are developed, based on heuristic methods for the serial assembly line worker assignment and balancing problem. Computational tests on literature instances and a detailed analysis of the results are reported
|
40 |
En jämförelse mellan klassisk webbprogrammering och no code : Fallstudie med företaget LyxrummetSharifzadeh, Shawbo January 2021 (has links)
Syftet med detta projekt har varit att skapa en jämförelse mellan två olika webbutvecklingsmetoder, no code och klassisk webbprogrammering, jämförelsen har skapats via en fallstudie hos företaget Lyxrummet där två webbutiker har skapats med dessa två olika metoder. No code-plattformen som använts har varit Wordpress och den klassiska webbprogrammerings metoden som använts har varit med språken PHP, HTML och CSS. För att kunna få fram resultatet av jämförelsen mellan webbutikerna så har användartester med 8 deltagare utförts samt mätningar på utvecklingstiden och kodrader. Slut resultatet visade att klassisk webbprogrammering var en mer lockande sida för deltagarna från användartestet men no code webbutiken var en snabbare och billigare variant för utveckling då ingen teknisk kunskap krävs. / The purpose of this project has been to create a comparison between the two different web development methods, no code and classical web development. The comparing has been done through a case study with the company Lyxrummet where two web shops have been created with these two methods. The no code platform that has been used is WordPress who is a leading platform for web development, and for the classical web development the language PHP, HTML, and CSS has been used. To be able to reach a result by the comprising between the web shops user testing have been done of 8 participates and measurement of the development time and code lines. The finishing result showed that the classical web developed web shop was the most liked by the user testing, but the no code web shop was a faster and cheaper alternative for web development because no technical knowledge person is needed.
|
Page generated in 0.0942 seconds