Internal model control of nonlinear systems through the inversion of recurrent neural networks

Recurrent neural networks can be used for both the identification and control of nonlinear systems. This paper takes a previously derived set of theoretical results about recurrent neural networks and applies them to the task of providing internal model control for a nonlinear plant. Using the theoretical results, we show how an inverse controller can be produced from a neural network model of the plant, without the need to train an additional network to perform the inverse control.

Dynamic input/output linearization using recurrent neural networks

A dynamic recurrent neural network (DRNN) is used to input/output linearize a control affine system in the globally linearizing control (GLC) structure. The network is trained as a part of a closed loop that involves a PI controller, the goal is to use the network, as a dynamic feedback, to cancel the nonlinear terms of the plant. The stability of the configuration is guarantee if the network and the plant are asymptotically stable and the linearizing input is bounded.

Identification of nonlinear systems with a dynamic recurrent neural network

Two approaches are presented to calculate the weights for a Dynamic Recurrent Neural Network (DRNN) in order to identify the input-output dynamics of a class of nonlinear systems. The number of states of the identified network is constrained to be the same as the number of states of the plant.

Relative order defines a topology for recurrent networks

This paper uses techniques from control theory in the analysis of trained recurrent neural networks. Differential geometry is used as a framework, which allows the concept of relative order to be applied to neural networks. Any system possessing finite relative order has a left-inverse. Any recurrent network with finite relative order also has an inverse, which is shown to be a recurrent network.

Computing journey start times with recurrent traffic conditions

n this study, the authors discuss the effective usage of technology to solve the problem of deciding on journey start times for recurrent traffic conditions. The developed algorithm guides the vehicles to travel on more reliable routes that are not easily prone to congestion or travel delays, ensures that the start time is as late as possible to avoid the traveller waiting too long at their destination and attempts to minimise the travel time. Experiments show that in order to be more certain of reaching their destination on time, a traveller has to leave early and correspondingly arrive early, resulting in a large waiting time. The application developed here asks the user to set this certainty factor as per the task in hand, and computes the best start time and route.

Coupling between gamma-band power and cerebral blood volume during recurrent acute neocortical seizures

Characterization of neural and hemodynamic biomarkers of epileptic activity that can be measured using noninvasive techniques is fundamental to the accurate identification of the epileptogenic zone (EZ) in the clinical setting. Recently, oscillations at gamma-band frequencies and above (N30 Hz) have been suggested to provide valuable localizing information of the EZ and track cortical activation associated with epileptogenic processes. Although a tight coupling between gamma-band activity and hemodynamic-based signals has been consistently demonstrated in non-pathological conditions, very little is known about whether such a relationship is maintained in epilepsy and the laminar etiology of these signals. Confirmation of this relationship may elucidate the underpinnings of perfusion-based signals in epilepsy and the potential value of localizing the EZ using hemodynamic correlates of pathological rhythms. Here, we use concurrent multi-depth electrophysiology and 2- dimensional optical imaging spectroscopy to examine the coupling between multi-band neural activity and cerebral blood volume (CBV) during recurrent acute focal neocortical seizures in the urethane-anesthetized rat. We show a powerful correlation between gamma-band power (25–90 Hz) and CBV across cortical laminae, in particular layer 5, and a close association between gamma measures and multi-unit activity (MUA). Our findings provide insights into the laminar electrophysiological basis of perfusion-based imaging signals in the epileptic state and may have implications for further research using non-invasive multi-modal techniques to localize epileptogenic tissue

Recurrent sores by ill-fitting dentures and intra-oral squamous cell carcinoma in smokers

Objectives: To examine whether denture use and recurrent sores caused by ill-fitting dentures are associated with intra-oral squamous cell carcinoma (IO-SCC) in individuals exposed to tobacco. Methods: We conducted a hospital-based case-control study. The study population comprised 124 patients with IO-SCC and the same number of controls (individually paired according to gender and age) recruited from outpatient units of the same hospital. Conditional logistic regression analysis assessed the effect of denture use and recurrent oral sores by ill-fitting dentures, adjusted by covariates on the lifetime exposure to alcohol and tobacco, socioeconomic standings, and dietary patterns. Results: The use of dentures showed no association with IO-SCC [adjusted odds ratio (OR) 1.40, 95 percent confidence interval 0.51-3.87, P = 0.513] in an assessment controlled by socioeconomic position, lifetime exposure to alcohol and tobacco, and dietary patterns. However, the report of recurrent sores caused by ill-fitting dentures showed significant association with the disease (adjusted OR 4.58, 95 percent confidence interval 1.52-13.76, P = 0.007). Conclusions: The association between recurrent oral sores caused by ill-fitting dentures and squamous cell carcinoma of the mouth in smokers is in agreement with the hypothesis that the chronic physical irritation of oral mucosa contributes to the topical carcinogenic effect of tobacco, which must be taken into careful consideration in the planning of dental services for adults and the elderly.

Recurrent Deletion of ZNF630 at Xp11.23 Is Not Associated With Mental Retardation

ZNF630 is a member of the primate-specific Xp11 zinc finger gene cluster that consists of six closely related genes, of which ZNF41, ZNF81, and ZNF674 have been shown to be involved in mental retardation. This suggests that mutations of ZNF630 might influence cognitive function. Here, we detected 12 ZNF630 deletions in a total of 1,562 male patients with mental retardation from Brazil, USA, Australia, and Europe. The breakpoints were analyzed in 10 families, and in all cases they were located within two segmental duplications that share more than 99% sequence identity, indicating that the deletions resulted from non-allelic homologous recombination. In 2,121 healthy male controls, 10 ZNF630 deletions were identified. In total, there was a 1.6-fold higher frequency of this deletion in males with mental retardation as compared to controls, but this increase was not statistically significant (P-value = 0.174). Conversely, a 1.9-fold lower frequency of ZNF630 duplications was observed in patients, which was not significant either (P-value = 0.163). These data do not show that ZNF630 deletions or duplications are associated with mental retardation. (C) 2010 Wiley-Liss, Inc.

Recurrent Parasitemias and Population Dynamics of Plasmodium vivax Polymorphisms in Rural Amazonia

Clinical trials documented alarming post-treatment Plasmodium vivax recurrence rates caused by recrudescence of surviving asexual blood stages, relapse from hypnozoites, or new infections. Here we describe high rates of P vivax recurrence (26-40% 180 days after treatment) in two cohorts of rural Amazonians exposed to low levels of malaria transmission after a vivax malaria episode treated with chloroquine-primaquine. Microsatellite analysis of 28 paired acute infection and recurrence parasites showed only two pairs of identical haplotypes (consistent with recrudescences or reactivation of homologous hypnozoites) and four pairs of related haplotypes (sharing alleles at 11-13 of 14 microsatellites analyzed). Local isolates of P vivax were extraordinarily diverse and rarely shared the same haplotype, indicating that frequent recurrences did not favor the persistence or reappearance of clonal lineages of parasites in the Population. This fast haplotype replacement rate may represent the typical population dynamics Of neutral polymorphisms in parasites from low-endemicity areas.

Risk of recurrent stroke and death after first stroke in long-distance ski race participants

Background: Physical activity is of benefit for primary prevention of cardiovascular diseases, but it appears to increase the risk for atrial fibrillation. We aimed to study a cohort of patients following a first stroke in individuals with previous high physical activity, compare them to the general population with respect to recurrent stroke and death, and relate these to atrial fibrillation. Methods and results: From the participants of the Vasaloppet, the world's largest ski-race, and matched individuals from the general population (n=708 604), we identified 5964 patients hospitalized with a first-time stroke between 1994 and 2010. Individuals with severe diseases were excluded. One half percent of skiers and 1% of nonskiers were hospitalized due to stroke. The incidence rate was 8.3 per 100 person-years among skiers and 11.1 among nonskiers. The hazard ratio (HR) for recurrent stroke or death between the 2 groups was 0.76 (95% CI 0.67 to 0.86). The result was consistent in subgroups. The HR for death was 0.66 (95% CI 0.56 to 0.78) and for recurrent stroke 0.82 (95% CI 0.70 to 0.96). After adjustment for smoking and socioeconomic factors, the HR for death was consistent at 0.70 (95% CI 0.56 to 0.87) while the HR for recurrent stroke was not statistically significant. Outcomes for skiers with atrial fibrillation tended to show a lower risk than for nonskiers. Conclusions: This large cohort study supports the hypothesis that patients with a stroke and with prior regular physical activity have a lower risk of death, while their risk for recurrent stroke is similar to that of nonskiers. The skiers had a higher incidence of atrial fibrillation, but still no increased risk of recurring stroke.

A novel approach based on recurrent neural networks applied to nonlinear systems optimization

This paper presents an efficient approach based on recurrent neural network for solving nonlinear optimization. More specifically, a modified Hopfield network is developed and its internal parameters are computed using the valid subspace technique. These parameters guarantee the convergence of the network to the equilibrium points that represent an optimal feasible solution. The main advantage of the developed network is that it treats optimization and constraint terms in different stages with no interference with each other. Moreover, the proposed approach does not require specification of penalty and weighting parameters for its initialization. A study of the modified Hopfield model is also developed to analyze its stability and convergence. Simulation results are provided to demonstrate the performance of the proposed neural network. (c) 2005 Elsevier B.V. All rights reserved.

Recurrent copy number gains of ACVR1 and corresponding transcript overexpression are associated with survival in head and neck squamous cell carcinomas

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

New recurrent deletions in the PPAR gamma and TP53 genes are associated with childhood myelodysplastic syndrome

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

Transcriptional changes associated with recurrent airway obstruction in affected and unaffected horses

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)