Inteligência artificial aplicada na previsão de mercados financeiros

Neste documento, são investigados vários métodos usados na inteligência artificial, com o objetivo de obter previsões precisas da evolução dos mercados financeiros. O uso de ferramentas lineares como os modelos AR, MA, ARMA e GARCH têm muitas limitações, pois torna-se muito difícil adaptá-los às não linearidades dos fenómenos que ocorrem nos mercados. Pelas razões anteriormente referidas, os algoritmos como as redes neuronais dinâmicas (TDNN, NARX e ESN), mostram uma maior capacidade de adaptação a estas não linearidades, pois não fazem qualquer pressuposto sobre as distribuições de probabilidade que caracterizam estes mercados. O facto destas redes neuronais serem dinâmicas, faz com que estas exibam um desempenho superior em relação às redes neuronais estáticas, ou outros algoritmos que não possuem qualquer tipo de memória. Apesar das vantagens reveladas pelas redes neuronais, estas são um sistema do tipo black box, o que torna muito difícil extrair informação dos pesos da rede. Isto significa que estes algoritmos devem ser usados com precaução, pois podem tornar-se instáveis.

Advances in scaling deep learning algorithms

Les algorithmes d'apprentissage profond forment un nouvel ensemble de méthodes puissantes pour l'apprentissage automatique. L'idée est de combiner des couches de facteurs latents en hierarchies. Cela requiert souvent un coût computationel plus elevé et augmente aussi le nombre de paramètres du modèle. Ainsi, l'utilisation de ces méthodes sur des problèmes à plus grande échelle demande de réduire leur coût et aussi d'améliorer leur régularisation et leur optimization. Cette thèse adresse cette question sur ces trois perspectives. Nous étudions tout d'abord le problème de réduire le coût de certains algorithmes profonds. Nous proposons deux méthodes pour entrainer des machines de Boltzmann restreintes et des auto-encodeurs débruitants sur des distributions sparses à haute dimension. Ceci est important pour l'application de ces algorithmes pour le traitement de langues naturelles. Ces deux méthodes (Dauphin et al., 2011; Dauphin and Bengio, 2013) utilisent l'échantillonage par importance pour échantilloner l'objectif de ces modèles. Nous observons que cela réduit significativement le temps d'entrainement. L'accéleration atteint 2 ordres de magnitude sur plusieurs bancs d'essai. Deuxièmement, nous introduisont un puissant régularisateur pour les méthodes profondes. Les résultats expérimentaux démontrent qu'un bon régularisateur est crucial pour obtenir de bonnes performances avec des gros réseaux (Hinton et al., 2012). Dans Rifai et al. (2011), nous proposons un nouveau régularisateur qui combine l'apprentissage non-supervisé et la propagation de tangente (Simard et al., 1992). Cette méthode exploite des principes géometriques et permit au moment de la publication d'atteindre des résultats à l'état de l'art. Finalement, nous considérons le problème d'optimiser des surfaces non-convexes à haute dimensionalité comme celle des réseaux de neurones. Tradionellement, l'abondance de minimum locaux était considéré comme la principale difficulté dans ces problèmes. Dans Dauphin et al. (2014a) nous argumentons à partir de résultats en statistique physique, de la théorie des matrices aléatoires, de la théorie des réseaux de neurones et à partir de résultats expérimentaux qu'une difficulté plus profonde provient de la prolifération de points-selle. Dans ce papier nous proposons aussi une nouvelle méthode pour l'optimisation non-convexe.

FPGA-based object detection using classification circuits

Dans l'apprentissage machine, la classification est le processus d’assigner une nouvelle observation à une certaine catégorie. Les classifieurs qui mettent en œuvre des algorithmes de classification ont été largement étudié au cours des dernières décennies. Les classifieurs traditionnels sont basés sur des algorithmes tels que le SVM et les réseaux de neurones, et sont généralement exécutés par des logiciels sur CPUs qui fait que le système souffre d’un manque de performance et d’une forte consommation d'énergie. Bien que les GPUs puissent être utilisés pour accélérer le calcul de certains classifieurs, leur grande consommation de puissance empêche la technologie d'être mise en œuvre sur des appareils portables tels que les systèmes embarqués. Pour rendre le système de classification plus léger, les classifieurs devraient être capable de fonctionner sur un système matériel plus compact au lieu d'un groupe de CPUs ou GPUs, et les classifieurs eux-mêmes devraient être optimisés pour ce matériel. Dans ce mémoire, nous explorons la mise en œuvre d'un classifieur novateur sur une plate-forme matérielle à base de FPGA. Le classifieur, conçu par Alain Tapp (Université de Montréal), est basé sur une grande quantité de tables de recherche qui forment des circuits arborescents qui effectuent les tâches de classification. Le FPGA semble être un élément fait sur mesure pour mettre en œuvre ce classifieur avec ses riches ressources de tables de recherche et l'architecture à parallélisme élevé. Notre travail montre que les FPGAs peuvent implémenter plusieurs classifieurs et faire les classification sur des images haute définition à une vitesse très élevée.

Los conceptos de salud financiera, riesgo y epidemiología en los estados financieros de las compañías del sector de extracción de petróleo crudo y gas natural en Colombia

El presente proyecto tiene como objeto identificar cuáles son los conceptos de salud, enfermedad, epidemiología y riesgo aplicables a las empresas del sector de extracción de petróleo y gas natural en Colombia. Dado, el bajo nivel de predicción de los análisis financieros tradicionales y su insuficiencia, en términos de inversión y toma de decisiones a largo plazo, además de no considerar variables como el riesgo y las expectativas de futuro, surge la necesidad de abordar diferentes perspectivas y modelos integradores. Esta apreciación es pertinente dentro del sector de extracción de petróleo y gas natural, debido a la creciente inversión extranjera que ha reportado, US$2.862 millones en el 2010, cifra mayor a diez veces su valor en el año 2003. Así pues, se podrían desarrollar modelos multi-dimensional, con base en los conceptos de salud financiera, epidemiológicos y estadísticos. El termino de salud y su adopción en el sector empresarial, resulta útil y mantiene una coherencia conceptual, evidenciando una presencia de diferentes subsistemas o factores interactuantes e interconectados. Es necesario mencionar también, que un modelo multidimensional (multi-stage) debe tener en cuenta el riesgo y el análisis epidemiológico ha demostrado ser útil al momento de determinarlo e integrarlo en el sistema junto a otros conceptos, como la razón de riesgo y riesgo relativo. Esto se analizará mediante un estudio teórico-conceptual, que complementa un estudio previo, para contribuir al proyecto de finanzas corporativas de la línea de investigación en Gerencia.

Synapsing variable length crossover: Biologically inspired crossover for variable length genomes

Biological Crossover occurs during the early stages of meiosis. During this process the chromosomes undergoing crossover are synapsed together at a number of homogenous sequence sections, it is within such synapsed sections that crossover occurs. The SVLC (Synapsing Variable Length Crossover) Algorithm recurrently synapses homogenous genetic sequences together in order of length. The genomes are considered to be flexible with crossover only being permitted within the synapsed sections. Consequently, common sequences are automatically preserved with only the genetic differences being exchanged, independent of the length of such differences. In addition to providing a rationale for variable length crossover it also provides a genotypic similarity metric for variable length genomes enabling standard niche formation techniques to be utilised. In a simple variable length test problem the SVLC algorithm outperforms current variable length crossover techniques.

Simple adaptive momentum: new algorithm for training multilayer perceptrons

The speed of convergence while training is an important consideration in the use of neural nets. The authors outline a new training algorithm which reduces both the number of iterations and training time required for convergence of multilayer perceptrons, compared to standard back-propagation and conjugate gradient descent algorithms.

A novel approach for distributed application scheduling based on prediction of communication events

The evolution of commodity computing lead to the possibility of efficient usage of interconnected machines to solve computationally-intensive tasks, which were previously solvable only by using expensive supercomputers. This, however, required new methods for process scheduling and distribution, considering the network latency, communication cost, heterogeneous environments and distributed computing constraints. An efficient distribution of processes over such environments requires an adequate scheduling strategy, as the cost of inefficient process allocation is unacceptably high. Therefore, a knowledge and prediction of application behavior is essential to perform effective scheduling. In this paper, we overview the evolution of scheduling approaches, focusing on distributed environments. We also evaluate the current approaches for process behavior extraction and prediction, aiming at selecting an adequate technique for online prediction of application execution. Based on this evaluation, we propose a novel model for application behavior prediction, considering chaotic properties of such behavior and the automatic detection of critical execution points. The proposed model is applied and evaluated for process scheduling in cluster and grid computing environments. The obtained results demonstrate that prediction of the process behavior is essential for efficient scheduling in large-scale and heterogeneous distributed environments, outperforming conventional scheduling policies by a factor of 10, and even more in some cases. Furthermore, the proposed approach proves to be efficient for online predictions due to its low computational cost and good precision. (C) 2009 Elsevier B.V. All rights reserved.

Machine learning and small robot navigation

Developing successful navigation and mapping strategies is an essential part of autonomous robot research. However, hardware limitations often make for inaccurate systems. This project serves to investigate efficient alternatives to mapping an environment, by first creating a mobile robot, and then applying machine learning to the robot and controlling systems to increase the robustness of the robot system. My mapping system consists of a semi-autonomous robot drone in communication with a stationary Linux computer system. There are learning systems running on both the robot and the more powerful Linux system. The first stage of this project was devoted to designing and building an inexpensive robot. Utilizing my prior experience from independent studies in robotics, I designed a small mobile robot that was well suited for simple navigation and mapping research. When the major components of the robot base were designed, I began to implement my design. This involved physically constructing the base of the robot, as well as researching and acquiring components such as sensors. Implementing the more complex sensors became a time-consuming task, involving much research and assistance from a variety of sources. A concurrent stage of the project involved researching and experimenting with different types of machine learning systems. I finally settled on using neural networks as the machine learning system to incorporate into my project. Neural nets can be thought of as a structure of interconnected nodes, through which information filters. The type of neural net that I chose to use is a type that requires a known set of data that serves to train the net to produce the desired output. Neural nets are particularly well suited for use with robotic systems as they can handle cases that lie at the extreme edges of the training set, such as may be produced by "noisy" sensor data. Through experimenting with available neural net code, I became familiar with the code and its function, and modified it to be more generic and reusable for multiple applications of neural nets.

Estudos de algumas ferramentas de coleta e visualização de dados e desempenho de aplicações paralelas no ambiente MPI

The last years have presented an increase in the acceptance and adoption of the parallel processing, as much for scientific computation of high performance as for applications of general intention. This acceptance has been favored mainly for the development of environments with massive parallel processing (MPP - Massively Parallel Processing) and of the distributed computation. A common point between distributed systems and MPPs architectures is the notion of message exchange, that allows the communication between processes. An environment of message exchange consists basically of a communication library that, acting as an extension of the programming languages that allow to the elaboration of applications parallel, such as C, C++ and Fortran. In the development of applications parallel, a basic aspect is on to the analysis of performance of the same ones. Several can be the metric ones used in this analysis: time of execution, efficiency in the use of the processing elements, scalability of the application with respect to the increase in the number of processors or to the increase of the instance of the treat problem. The establishment of models or mechanisms that allow this analysis can be a task sufficiently complicated considering parameters and involved degrees of freedom in the implementation of the parallel application. An joined alternative has been the use of collection tools and visualization of performance data, that allow the user to identify to points of strangulation and sources of inefficiency in an application. For an efficient visualization one becomes necessary to identify and to collect given relative to the execution of the application, stage this called instrumentation. In this work it is presented, initially, a study of the main techniques used in the collection of the performance data, and after that a detailed analysis of the main available tools is made that can be used in architectures parallel of the type to cluster Beowulf with Linux on X86 platform being used libraries of communication based in applications MPI - Message Passing Interface, such as LAM and MPICH. This analysis is validated on applications parallel bars that deal with the problems of the training of neural nets of the type perceptrons using retro-propagation. The gotten conclusions show to the potentiality and easinesses of the analyzed tools.

Human face verification based on multidimensional Polynomial Powers of Sigmoid (PPS)

In this paper, we described how a multidimensional wavelet neural networks based on Polynomial Powers of Sigmoid (PPS) can be constructed, trained and applied in image processing tasks. In this sense, a novel and uniform framework for face verification is presented. The framework is based on a family of PPS wavelets,generated from linear combination of the sigmoid functions, and can be considered appearance based in that features are extracted from the face image. The feature vectors are then subjected to subspace projection of PPS-wavelet. The design of PPS-wavelet neural networks is also discussed, which is seldom reported in the literature. The Stirling Universitys face database were used to generate the results. Our method has achieved 92 % of correct detection and 5 % of false detection rate on the database.

A novel intelligent system to nitrogen content prediction in plants using indirect chlorophyll measurements

The accurate identification of the nitrogen content in plants is extremely important since it involves economic aspects and environmental impacts, Several experimental tests have been carried out to obtain characteristics and parameters associated with the health of plants and its growing. The nitrogen content identification in plants involves a lot of non-linear parameters and complexes mathematical models. This paper describes a novel approach for identification of nitrogen content thought SPAD index using artificial neural networks (ANN). The network acts as identifier of relationships among, crop varieties, fertilizer treatments, type of leaf and nitrogen content in the plants (target). So, nitrogen content can be generalized and estimated and from an input parameter set. This approach can form the basis for development of an accurate real time system to predict nitrogen content in plants.

A fuzzy logic system applied in lightning models

This paper describes a novel approach for mapping lightning processes using fuzzy logic. The estimation process is carried out using a fuzzy system based on Sugeno's architecture. Simulation results confirm that proposed approach can be efficiently used in these types of problem.

Estudo da predição da circularidade e rugosidade de peças retificadas utilizando as redes neurais artificiais

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)

Localização de falhas em linhas de transmissão utilizando decomposição harmônica e redes neurais artificiais

Este trabalho propõe a utilização de uma nova metodologia para a localização de falhas em linhas de transmissão (LT). Esta metodologia consiste na utilização da decomposição harmônica da corrente de fuga de uma linha e na aplicação de uma Rede Neural Artificial (RNA) capaz de distinguir padrões da condição normal de funcionamento e padrões de situações de falhas de uma LT. Foi utilizado um modelo Pi capaz de absorver dados reais de tensão e corrente de três fases e de alterar valores de R, L e C segundo modificações ambientais. Neste modelo foram geradas falhas em todas as torres com diferentes valores de capacitância. A saída fornecida pelo modelo é a decomposição da corrente de fuga do trecho considerado. Os dados de entrada e saída do modelo foram utilizados no treinamento da RNA desenvolvida. A aquisição de dados reais de tensão e corrente foi feita através de analisadores de parâmetros de qualidade de energia elétrica instalados nas extremidades de um trecho de LT, Guamá-Utinga, pertencente à Centrais Elétricas do Norte do Brasil ELETRONORTE. O cálculo dos parâmetros construtivos foi feito através do método matricial e melhorado através da utilização do Método de Elementos Finitos (MEF). A RNA foi desenvolvida com o auxílio do software Matlab. Para treinamento da RNA foi utilizado o algoritmo de Retropropagação Resiliente que apresentou um bom desempenho. A RNA foi treinada com dois conjuntos de dados de treinamento para analisar possíveis diferenças entre as saídas fornecidas pelos dois grupos. Nos dois casos apresentou resultados satisfatórios, possibilitando a localização de falhas no trecho considerado.