2018 17th IEEE International Conference on Machine Learning and Applications (ICMLA), 2018
Orthogonal regression extends the classical regression framework by assuming that the data may co... more Orthogonal regression extends the classical regression framework by assuming that the data may contain errors in both the dependent and independent variables. Often, this approach tends to outperform classical regression in real-world scenarios. However, the algorithms used to determine a solution to the orthogonal regression problem require the computation of singular value decompositions (SVD), which may be computationally expensive and impractical for real-world problems. In this work, we propose a new approach to the orthogonal regression problem based on a regularized squared loss. The method follows an online learning strategy which makes it more flexible for different types of applications. The algorithm is derived in primal and dual variables and the later formulation allows the introduction of kernels for nonlinear modeling. We compare our proposed orthogonal regression algorithm to a corresponding classical regression algorithm using both synthetic and real-world datasets from different applications. Our algorithm achieved better results for most of the datasets.
A demanda por sistemas de comunicacao de alto desempenho, que necessitam operar de forma muito ef... more A demanda por sistemas de comunicacao de alto desempenho, que necessitam operar de forma muito eficiente, cria novos desafios em sua modelagem. A analise via Trafego Pesado estuda justamente modelos de sistemas de filas que operam no limite de sua saturacao. Atraves desta tecnica, prova-se que no limite da saturacao do sistema, detalhes minuciosos aparecem como ruidos, descritos por um processo de movimento Browniano, e as caracteristicas essenciais sao colocadas em evidencia, como termos de deriva. Desta forma, e possivel descrever o sistema utilizando somente os dois primeiros momentos dos processos estocasticos envolvidos. Nestes casos, pode-se construir controles otimos utilizando o Metodo da Cadeia de Markov Aproximada. Estes metodos baseiam-se em uma discretizacao do espaco de estados para tratar o problema de controle como um de Processo de Decisao de Markov, que podem ser resolvidos atraves dos chamados metodos de Programacao Dinâmica.
In this paper, we introduce a new online algorithm for orthogonal regression. The method is const... more In this paper, we introduce a new online algorithm for orthogonal regression. The method is constructed via a stochastic gradient descent approach combined with the idea of a tube loss function, which is similar to the one used in support vector (SV) regression. The algorithm can be used in primal or in dual variables. The latter formulation allows the introduction of kernels and soft margins. In addition, an incremental strategy algorithm is introduced, which can be used to find sparse solutions and also an approximation to the “minimal tube” containing the data. The algorithm is very simple to implement and avoids quadratic optimization.
Este trabalho discute a aplicacao de tecnicas de aprendizado por reforco no âmbito de futebol de ... more Este trabalho discute a aplicacao de tecnicas de aprendizado por reforco no âmbito de futebol de robos. Por se tratar de um problema de natureza continua, requer metodos especiais para aproximacoes de funcoes.
In this paper, we investigate optimal power management in parallel processing systems composed of... more In this paper, we investigate optimal power management in parallel processing systems composed of one queue and several identical processing stations. Power consumption is controlled by setting some of the stations into an inactive state, where they consume less power but are unable to provide service. This way, we are faced with the conflicting objective of minimizing power consumption while maintaining a desired quality of service. A distinguishing feature here, regarding most previous literature on this subject, is that we consider systems operating a policy that may turn the reserve machines on or off with setup times and under general inter-arrival or service time distributions, subject to some conditions. When these condition fail, we also provide a model with general inter-arrival times and exponentially distributed service times. To some extent, a controlled switching diffusion obtained in this paper via heavy traffic analysis and stochastic optimal control theory are the te...
Microarray experiments are capable of measuring the expression level of thousands of genes simult... more Microarray experiments are capable of measuring the expression level of thousands of genes simultaneously. Dealing with this enormous amount of information requires complex computation. Support Vector Machines (SVM) have been widely used with great efficiency to solve classification problems that have high dimension. In this sense, it is plausible to develop new feature selection strategies for microarray data that are associated with this type of classifier. Therefore, we propose, in this paper, a new method for feature selection based on an ordered search process to explore the space of possible subsets. The algorithm, called Admissible Ordered Search (AOS), uses as evaluation function the margin values estimated for each hypothesis by a SVM classifier. An important theoretical contribution of this paper is the development of the projected margin concept. This value is computed as the margin vector projection on a lower dimensional subspace and is used as an upper bound for the cu...
Advances in Bioinformatics and Computational Biology, 2020
The increasing amount of data available on the web, coupled with the demand for useful informatio... more The increasing amount of data available on the web, coupled with the demand for useful information, has sparked increasing interest in gaining knowledge in large information systems, especially biomedical ones. Health institutions operate in an environment that has been generating thousands of health records about patients. Such databases can be the source of a wealth of information. For instance, these databases can be used to study factors that contribute to the incidence of a pathology and thereby determine patient profiles at the earliest stage of the disease. Such information can be extracted with the help of Machine Learning methods, which are capable of dealing with large amounts of data in order to make predictions. These methods offer an opportunity to translate new data into palpable information and, thus, allows earlier diagnosis and precise treatment options. In order to understand the potential of these methods, we use a database that contain records of cancer patients, which is made publicly available by the Oncocentro Foundation of Sao Paulo. This database contains historical clinical information from cancer patients of the past 20 years. In this paper we present an initial investigation towards the goal of improving prognosis and therefore increasing the chances of survival among cancer patients. The Random Forest Classification Model was employed in our analysis; this model shows to be a suitable predicting tool for ours purpose. Thus, we intend to present means that allows the design of predictive, preventive and personalized treatments, as well as assisting in the decision making process of the disease.
In many cell types, epigenetic changes are partially regulated by the availability of metabolites... more In many cell types, epigenetic changes are partially regulated by the availability of metabolites involved in the activity of chromatin-modifying enzymes. Even so, the association between metabolism and the typical epigenetic reprogramming that occurs during preimplantation embryo development remains poorly understood. In this work, we explore the link between energy metabolism, more specifically the tricarboxylic acid cycle (TCA), and epigenetic regulation in bovine preimplantation embryos. Using a morphokinetics model of embryonic development (fast- and slow-developing embryos), we show that DNA methylation (5mC) and hydroxymethylation (5hmC) are dynamically regulated and altered by the speed of the first cleavages. More specifically, slow-developing embryos fail to perform the typical reprogramming that is necessary to ensure the generation of blastocysts with higher ability to establish specific cell lineages. Transcriptome analysis revealed that such differences were mainly ass...
Abstract Repeated games is a branch of game theory, where a game can be played several times by t... more Abstract Repeated games is a branch of game theory, where a game can be played several times by the players involved. In this setting, players may not always play the optimal strategy or they may be willing to engage in collaboration or other types of behavior which might lead to a higher long-term profit. Since the same game is repeated for several rounds, and considering a scenario with complete information, it is possible for a player to analyze its opponent’s behavior in order to find patterns. These patterns can then be used to predict the opponent’s actions. Such a setting, where players have mutual information about past moves and do not always play in equilibrium, leads naturally to non-stationary environments, where the players can frequently modify their strategies in order to get ahead in the game. In this work, we propose a novel algorithm based on a string kernel density estimation, which is capable of predicting the opponent’s actions in repeated games and can be used to optimize the player’s profit over time. The prediction is not limited to the next round action. It can also be used to predict a finite sequence of future rounds, which can be combined with a lookahead search scheme with limited depth. In the experiments section, it is shown that the proposed algorithm is able to learn and adapt rapidly, providing good results even if the opponent also adopts an adaptive strategy.
RESUMOObjetivo: descrever os principais motivos que levaram os idosos à institucionalização. Méto... more RESUMOObjetivo: descrever os principais motivos que levaram os idosos à institucionalização. Método: estudo quantitativo, transversal, realizado em uma instituição de longa permanência para idosos, por consulta a 219 prontuários, utilizando-se um questionário. Para a associação das variáveis contínuas e categóricas, foram utilizados os testes Kruskal-wallis e o Teste Qui-Quadrado de Pearson considerando p<0,05, apresentados em tabela. Resultados: no total, 55,6% eram idosos do sexo feminino, com idade média de 77 (±0,55). Os principais motivos foram vontade própria, questões familiares, abandono e violência. As variáveis mais relacionadas aos motivos de institucionalização foram sexo (p=0,013), estado civil (p=0,041), com quem residia (p=<0,001), recebe visitas (p=0,011) e grau de dependência (p=<0,001). As causas determinantes encontradas que levaram os idosos à inserção em uma instituição de longa permanência chamam a atenção para a busca de ações e estratégias sociais e ...
Continuous-time Markov chain models are often used to describe the stochastic dynamics of network... more Continuous-time Markov chain models are often used to describe the stochastic dynamics of networks of reacting chemical species, especially in the growing field of systems biology. These Markov chain models are often studied by simulating sample paths in order to generate Monte-Carlo estimates. However, discrete-event stochastic simulation of these models rapidly becomes computationally intensive. Consequently, more tractable diffusion approximations are commonly used in numerical computation, even for modest-sized networks. However, existing approximations either do not respect the constraint that chemical concentrations are never negative (linear noise approximation) or are typically only valid until the concentration of some chemical species first becomes zero (Langevin approximation). In this paper, we propose an approximation for such Markov chains via reflected diffusion processes that respect the fact that concentrations of chemical species are never negative. We call this a constrained Langevin approximation because it behaves like the Langevin approximation in the interior of the positive orthant, to which it is constrained by instantaneous reflection at the boundary of the orthant. An additional advantage of our approximation is that it can be written down immediately from the chemical reactions. This contrasts with the linear noise approximation, which involves a twostage procedure-first solve a deterministic reaction rate ordinary differential equation, followed by a stochastic differential equation for fluctuations around those solutions. Our approximation also captures the interaction of nonlinearities in the reaction rate function with the driving noise. In simulations, we have found the computation time for our approximation to be at least comparable to, and often better than, that for the linear noise approximation. Under mild assumptions, we first prove that our proposed approximation is well defined for all time. Then we prove that it can be obtained as the weak limit of a sequence of jump-diffusion processes that behave like the Langevin approximation in the interior of the positive orthant and like a rescaled version of the Markov chain on the boundary of the orthant. For this limit the
Stochastic effects play an important role in modeling the time evolution of chemical reaction sys... more Stochastic effects play an important role in modeling the time evolution of chemical reaction systems in fields such as systems biology, where the concentrations of some constituent molecules can be low. The most common stochastic models for these systems are continuous time Markov chains, which track the molecular abundance of each chemical species. Often, these stochastic models are studied by computer simulations, which can quickly become computationally expensive. A common approach to reduce computational effort is to approximate the discrete valued Markov chain by a continuous valued diffusion process. However, existing diffusion approximations either do not respect the constraint that chemical concentrations are never negative (linear noise approximation) or are typically only valid until the concentration of some chemical species first becomes zero (chemical Langevin equation). In this paper, we propose (obliquely) reflected diffusions, which respect the nonnegativity of chemical concentrations, as approximations for Markov chain models of chemical reaction networks. These reflected diffusions satisfy``constrained Langevin equations,"" in that they behave like solutions of chemical Langevin equations in the interior of the positive orthant and are constrained to the orthant by instantaneous oblique reflection at the boundary. To motivate their form, we first illustrate our constrained Langevin approximations for two simple examples. We then describe the general form of our proposed approximation. We illustrate the performance of our approximations through comparison of their stationary distributions for the two examples with those of the Markov chain model and through simulations of more complex examples.
Anais do 9. Congresso Brasileiro de Redes Neurais, 2016
Resumo-Este trabalho tem como objetivo apresentar uma nova formulação para o problema de obtenção... more Resumo-Este trabalho tem como objetivo apresentar uma nova formulação para o problema de obtenção de um classificador de máxima margem com norma arbitrária. O trabalho apresenta, também, um novo algoritmo baseado na solução de sucessivos sistemas de inequações com o uso de uma versão modificada do algoritmo de treinamento do perceptron. A sua grande vantagem consiste na sua flexibilidade, que permite a utilização de diferentes valores de norma, ao contrário de algumas formulações especiais existentes que se restringem à utilização das normas L 1 e L ∞. O algoritmo foi chamado de IMA p pois se trata de um caso especial do algoritmo IMA desenvolvido para a obtenção de classificadores de máxima margem com o uso da norma euclidiana.
Introduction: Nursing care for institutionalized elderly should be done through systematized acti... more Introduction: Nursing care for institutionalized elderly should be done through systematized actions and by using the nursing process, directing the care actions to the affected needs with theoretical support and use of classification systems. In this context, the objective was to develop nursing diagnoses of the International Classification for Nursing Practice for institutionalized elderly. Method: This is a descriptive study, developed in a Long-stay Institution for Elderly People in Northeast Brazil. The survey was conducted from April to May 2016, with 28 institutionalized elderly. Data collection was guided by a form based in Henderson's theory and carried out through physical examination, clinical interview and records consultation. The nursing diagnoses were developed by using the International Classification for Nursing Practice version 2015. Data were analyzed using SPSS version 20.0. Results: Authors found 81 types of diagnoses, among which there was highlight to: impaired heart condition, productive cough, obesity, urinary incontinence, constipation, diarrhea, risk for falls, impaired sleep, impaired ability to perform hygiene, edema in the lower limbs and impaired vision, lack of recreational activity, positive socialization and positive religious belief. Most nursing diagnoses belonged to the needs of moving and maintaining desirable posture and communicating. Conclusion: There was an increased frequency of diagnoses related to biological needs.
Anais do 11. Congresso Brasileiro de Inteligência Computacional, 2016
In this contribution, we introduce an online method for regression composed of two parts. The fir... more In this contribution, we introduce an online method for regression composed of two parts. The first is based on a stochastic gradient descent approach combined with the idea of tube used in support vector regression. This algorithm can be used in primal or in dual variables. The latter formulation allows the introduction of kernels and soft margins. The second part consists of an incremental strategy algorithm which is introduced in order to find sparse solutions. Also, when soft margin is not desirable, this incremental strategy may be used to obtain the "minimal tube" containing the data. The algorithm is very simple to implement and avoids quadratic optimization. Numerical results show that the method works well in comparison to a standard implementation of the SV-regression.
... Saul C. Leite a , E-mail The Corresponding Author and Marcelo D. Fragoso Corresponding Author... more ... Saul C. Leite a , E-mail The Corresponding Author and Marcelo D. Fragoso Corresponding Author Contact Information , a , E-mail The Corresponding Author [Author vitae]. a Department of Systems and Control, National Laboratory for Scientific Computing (LNCC), Av. ...
We consider the problem of reducing the response time of fork-join systems by maintaining the wor... more We consider the problem of reducing the response time of fork-join systems by maintaining the workload balanced among the processing stations. The general problem of modeling and finding an optimal policy that reduces imbalance is quite difficult. Hence, the heavy traffic approach is taken, and the system dynamics is approximated by a reflected diffusion process. This way, the problem of finding an optimal balancing policy that reduces workload imbalance is set as an stochastic optimal control problem, for which numerical methods are available. Some numerical experiments are presented, where the control problem is solved numerically and applied to a simulation. The results indicate that the response time of the controlled system is reduced significantly using the devised control.
2018 17th IEEE International Conference on Machine Learning and Applications (ICMLA), 2018
Orthogonal regression extends the classical regression framework by assuming that the data may co... more Orthogonal regression extends the classical regression framework by assuming that the data may contain errors in both the dependent and independent variables. Often, this approach tends to outperform classical regression in real-world scenarios. However, the algorithms used to determine a solution to the orthogonal regression problem require the computation of singular value decompositions (SVD), which may be computationally expensive and impractical for real-world problems. In this work, we propose a new approach to the orthogonal regression problem based on a regularized squared loss. The method follows an online learning strategy which makes it more flexible for different types of applications. The algorithm is derived in primal and dual variables and the later formulation allows the introduction of kernels for nonlinear modeling. We compare our proposed orthogonal regression algorithm to a corresponding classical regression algorithm using both synthetic and real-world datasets from different applications. Our algorithm achieved better results for most of the datasets.
A demanda por sistemas de comunicacao de alto desempenho, que necessitam operar de forma muito ef... more A demanda por sistemas de comunicacao de alto desempenho, que necessitam operar de forma muito eficiente, cria novos desafios em sua modelagem. A analise via Trafego Pesado estuda justamente modelos de sistemas de filas que operam no limite de sua saturacao. Atraves desta tecnica, prova-se que no limite da saturacao do sistema, detalhes minuciosos aparecem como ruidos, descritos por um processo de movimento Browniano, e as caracteristicas essenciais sao colocadas em evidencia, como termos de deriva. Desta forma, e possivel descrever o sistema utilizando somente os dois primeiros momentos dos processos estocasticos envolvidos. Nestes casos, pode-se construir controles otimos utilizando o Metodo da Cadeia de Markov Aproximada. Estes metodos baseiam-se em uma discretizacao do espaco de estados para tratar o problema de controle como um de Processo de Decisao de Markov, que podem ser resolvidos atraves dos chamados metodos de Programacao Dinâmica.
In this paper, we introduce a new online algorithm for orthogonal regression. The method is const... more In this paper, we introduce a new online algorithm for orthogonal regression. The method is constructed via a stochastic gradient descent approach combined with the idea of a tube loss function, which is similar to the one used in support vector (SV) regression. The algorithm can be used in primal or in dual variables. The latter formulation allows the introduction of kernels and soft margins. In addition, an incremental strategy algorithm is introduced, which can be used to find sparse solutions and also an approximation to the “minimal tube” containing the data. The algorithm is very simple to implement and avoids quadratic optimization.
Este trabalho discute a aplicacao de tecnicas de aprendizado por reforco no âmbito de futebol de ... more Este trabalho discute a aplicacao de tecnicas de aprendizado por reforco no âmbito de futebol de robos. Por se tratar de um problema de natureza continua, requer metodos especiais para aproximacoes de funcoes.
In this paper, we investigate optimal power management in parallel processing systems composed of... more In this paper, we investigate optimal power management in parallel processing systems composed of one queue and several identical processing stations. Power consumption is controlled by setting some of the stations into an inactive state, where they consume less power but are unable to provide service. This way, we are faced with the conflicting objective of minimizing power consumption while maintaining a desired quality of service. A distinguishing feature here, regarding most previous literature on this subject, is that we consider systems operating a policy that may turn the reserve machines on or off with setup times and under general inter-arrival or service time distributions, subject to some conditions. When these condition fail, we also provide a model with general inter-arrival times and exponentially distributed service times. To some extent, a controlled switching diffusion obtained in this paper via heavy traffic analysis and stochastic optimal control theory are the te...
Microarray experiments are capable of measuring the expression level of thousands of genes simult... more Microarray experiments are capable of measuring the expression level of thousands of genes simultaneously. Dealing with this enormous amount of information requires complex computation. Support Vector Machines (SVM) have been widely used with great efficiency to solve classification problems that have high dimension. In this sense, it is plausible to develop new feature selection strategies for microarray data that are associated with this type of classifier. Therefore, we propose, in this paper, a new method for feature selection based on an ordered search process to explore the space of possible subsets. The algorithm, called Admissible Ordered Search (AOS), uses as evaluation function the margin values estimated for each hypothesis by a SVM classifier. An important theoretical contribution of this paper is the development of the projected margin concept. This value is computed as the margin vector projection on a lower dimensional subspace and is used as an upper bound for the cu...
Advances in Bioinformatics and Computational Biology, 2020
The increasing amount of data available on the web, coupled with the demand for useful informatio... more The increasing amount of data available on the web, coupled with the demand for useful information, has sparked increasing interest in gaining knowledge in large information systems, especially biomedical ones. Health institutions operate in an environment that has been generating thousands of health records about patients. Such databases can be the source of a wealth of information. For instance, these databases can be used to study factors that contribute to the incidence of a pathology and thereby determine patient profiles at the earliest stage of the disease. Such information can be extracted with the help of Machine Learning methods, which are capable of dealing with large amounts of data in order to make predictions. These methods offer an opportunity to translate new data into palpable information and, thus, allows earlier diagnosis and precise treatment options. In order to understand the potential of these methods, we use a database that contain records of cancer patients, which is made publicly available by the Oncocentro Foundation of Sao Paulo. This database contains historical clinical information from cancer patients of the past 20 years. In this paper we present an initial investigation towards the goal of improving prognosis and therefore increasing the chances of survival among cancer patients. The Random Forest Classification Model was employed in our analysis; this model shows to be a suitable predicting tool for ours purpose. Thus, we intend to present means that allows the design of predictive, preventive and personalized treatments, as well as assisting in the decision making process of the disease.
In many cell types, epigenetic changes are partially regulated by the availability of metabolites... more In many cell types, epigenetic changes are partially regulated by the availability of metabolites involved in the activity of chromatin-modifying enzymes. Even so, the association between metabolism and the typical epigenetic reprogramming that occurs during preimplantation embryo development remains poorly understood. In this work, we explore the link between energy metabolism, more specifically the tricarboxylic acid cycle (TCA), and epigenetic regulation in bovine preimplantation embryos. Using a morphokinetics model of embryonic development (fast- and slow-developing embryos), we show that DNA methylation (5mC) and hydroxymethylation (5hmC) are dynamically regulated and altered by the speed of the first cleavages. More specifically, slow-developing embryos fail to perform the typical reprogramming that is necessary to ensure the generation of blastocysts with higher ability to establish specific cell lineages. Transcriptome analysis revealed that such differences were mainly ass...
Abstract Repeated games is a branch of game theory, where a game can be played several times by t... more Abstract Repeated games is a branch of game theory, where a game can be played several times by the players involved. In this setting, players may not always play the optimal strategy or they may be willing to engage in collaboration or other types of behavior which might lead to a higher long-term profit. Since the same game is repeated for several rounds, and considering a scenario with complete information, it is possible for a player to analyze its opponent’s behavior in order to find patterns. These patterns can then be used to predict the opponent’s actions. Such a setting, where players have mutual information about past moves and do not always play in equilibrium, leads naturally to non-stationary environments, where the players can frequently modify their strategies in order to get ahead in the game. In this work, we propose a novel algorithm based on a string kernel density estimation, which is capable of predicting the opponent’s actions in repeated games and can be used to optimize the player’s profit over time. The prediction is not limited to the next round action. It can also be used to predict a finite sequence of future rounds, which can be combined with a lookahead search scheme with limited depth. In the experiments section, it is shown that the proposed algorithm is able to learn and adapt rapidly, providing good results even if the opponent also adopts an adaptive strategy.
RESUMOObjetivo: descrever os principais motivos que levaram os idosos à institucionalização. Méto... more RESUMOObjetivo: descrever os principais motivos que levaram os idosos à institucionalização. Método: estudo quantitativo, transversal, realizado em uma instituição de longa permanência para idosos, por consulta a 219 prontuários, utilizando-se um questionário. Para a associação das variáveis contínuas e categóricas, foram utilizados os testes Kruskal-wallis e o Teste Qui-Quadrado de Pearson considerando p<0,05, apresentados em tabela. Resultados: no total, 55,6% eram idosos do sexo feminino, com idade média de 77 (±0,55). Os principais motivos foram vontade própria, questões familiares, abandono e violência. As variáveis mais relacionadas aos motivos de institucionalização foram sexo (p=0,013), estado civil (p=0,041), com quem residia (p=<0,001), recebe visitas (p=0,011) e grau de dependência (p=<0,001). As causas determinantes encontradas que levaram os idosos à inserção em uma instituição de longa permanência chamam a atenção para a busca de ações e estratégias sociais e ...
Continuous-time Markov chain models are often used to describe the stochastic dynamics of network... more Continuous-time Markov chain models are often used to describe the stochastic dynamics of networks of reacting chemical species, especially in the growing field of systems biology. These Markov chain models are often studied by simulating sample paths in order to generate Monte-Carlo estimates. However, discrete-event stochastic simulation of these models rapidly becomes computationally intensive. Consequently, more tractable diffusion approximations are commonly used in numerical computation, even for modest-sized networks. However, existing approximations either do not respect the constraint that chemical concentrations are never negative (linear noise approximation) or are typically only valid until the concentration of some chemical species first becomes zero (Langevin approximation). In this paper, we propose an approximation for such Markov chains via reflected diffusion processes that respect the fact that concentrations of chemical species are never negative. We call this a constrained Langevin approximation because it behaves like the Langevin approximation in the interior of the positive orthant, to which it is constrained by instantaneous reflection at the boundary of the orthant. An additional advantage of our approximation is that it can be written down immediately from the chemical reactions. This contrasts with the linear noise approximation, which involves a twostage procedure-first solve a deterministic reaction rate ordinary differential equation, followed by a stochastic differential equation for fluctuations around those solutions. Our approximation also captures the interaction of nonlinearities in the reaction rate function with the driving noise. In simulations, we have found the computation time for our approximation to be at least comparable to, and often better than, that for the linear noise approximation. Under mild assumptions, we first prove that our proposed approximation is well defined for all time. Then we prove that it can be obtained as the weak limit of a sequence of jump-diffusion processes that behave like the Langevin approximation in the interior of the positive orthant and like a rescaled version of the Markov chain on the boundary of the orthant. For this limit the
Stochastic effects play an important role in modeling the time evolution of chemical reaction sys... more Stochastic effects play an important role in modeling the time evolution of chemical reaction systems in fields such as systems biology, where the concentrations of some constituent molecules can be low. The most common stochastic models for these systems are continuous time Markov chains, which track the molecular abundance of each chemical species. Often, these stochastic models are studied by computer simulations, which can quickly become computationally expensive. A common approach to reduce computational effort is to approximate the discrete valued Markov chain by a continuous valued diffusion process. However, existing diffusion approximations either do not respect the constraint that chemical concentrations are never negative (linear noise approximation) or are typically only valid until the concentration of some chemical species first becomes zero (chemical Langevin equation). In this paper, we propose (obliquely) reflected diffusions, which respect the nonnegativity of chemical concentrations, as approximations for Markov chain models of chemical reaction networks. These reflected diffusions satisfy``constrained Langevin equations,"" in that they behave like solutions of chemical Langevin equations in the interior of the positive orthant and are constrained to the orthant by instantaneous oblique reflection at the boundary. To motivate their form, we first illustrate our constrained Langevin approximations for two simple examples. We then describe the general form of our proposed approximation. We illustrate the performance of our approximations through comparison of their stationary distributions for the two examples with those of the Markov chain model and through simulations of more complex examples.
Anais do 9. Congresso Brasileiro de Redes Neurais, 2016
Resumo-Este trabalho tem como objetivo apresentar uma nova formulação para o problema de obtenção... more Resumo-Este trabalho tem como objetivo apresentar uma nova formulação para o problema de obtenção de um classificador de máxima margem com norma arbitrária. O trabalho apresenta, também, um novo algoritmo baseado na solução de sucessivos sistemas de inequações com o uso de uma versão modificada do algoritmo de treinamento do perceptron. A sua grande vantagem consiste na sua flexibilidade, que permite a utilização de diferentes valores de norma, ao contrário de algumas formulações especiais existentes que se restringem à utilização das normas L 1 e L ∞. O algoritmo foi chamado de IMA p pois se trata de um caso especial do algoritmo IMA desenvolvido para a obtenção de classificadores de máxima margem com o uso da norma euclidiana.
Introduction: Nursing care for institutionalized elderly should be done through systematized acti... more Introduction: Nursing care for institutionalized elderly should be done through systematized actions and by using the nursing process, directing the care actions to the affected needs with theoretical support and use of classification systems. In this context, the objective was to develop nursing diagnoses of the International Classification for Nursing Practice for institutionalized elderly. Method: This is a descriptive study, developed in a Long-stay Institution for Elderly People in Northeast Brazil. The survey was conducted from April to May 2016, with 28 institutionalized elderly. Data collection was guided by a form based in Henderson's theory and carried out through physical examination, clinical interview and records consultation. The nursing diagnoses were developed by using the International Classification for Nursing Practice version 2015. Data were analyzed using SPSS version 20.0. Results: Authors found 81 types of diagnoses, among which there was highlight to: impaired heart condition, productive cough, obesity, urinary incontinence, constipation, diarrhea, risk for falls, impaired sleep, impaired ability to perform hygiene, edema in the lower limbs and impaired vision, lack of recreational activity, positive socialization and positive religious belief. Most nursing diagnoses belonged to the needs of moving and maintaining desirable posture and communicating. Conclusion: There was an increased frequency of diagnoses related to biological needs.
Anais do 11. Congresso Brasileiro de Inteligência Computacional, 2016
In this contribution, we introduce an online method for regression composed of two parts. The fir... more In this contribution, we introduce an online method for regression composed of two parts. The first is based on a stochastic gradient descent approach combined with the idea of tube used in support vector regression. This algorithm can be used in primal or in dual variables. The latter formulation allows the introduction of kernels and soft margins. The second part consists of an incremental strategy algorithm which is introduced in order to find sparse solutions. Also, when soft margin is not desirable, this incremental strategy may be used to obtain the "minimal tube" containing the data. The algorithm is very simple to implement and avoids quadratic optimization. Numerical results show that the method works well in comparison to a standard implementation of the SV-regression.
... Saul C. Leite a , E-mail The Corresponding Author and Marcelo D. Fragoso Corresponding Author... more ... Saul C. Leite a , E-mail The Corresponding Author and Marcelo D. Fragoso Corresponding Author Contact Information , a , E-mail The Corresponding Author [Author vitae]. a Department of Systems and Control, National Laboratory for Scientific Computing (LNCC), Av. ...
We consider the problem of reducing the response time of fork-join systems by maintaining the wor... more We consider the problem of reducing the response time of fork-join systems by maintaining the workload balanced among the processing stations. The general problem of modeling and finding an optimal policy that reduces imbalance is quite difficult. Hence, the heavy traffic approach is taken, and the system dynamics is approximated by a reflected diffusion process. This way, the problem of finding an optimal balancing policy that reduces workload imbalance is set as an stochastic optimal control problem, for which numerical methods are available. Some numerical experiments are presented, where the control problem is solved numerically and applied to a simulation. The results indicate that the response time of the controlled system is reduced significantly using the devised control.
Uploads
Papers by Saul Leite