Nonlinear time series prediction based on a power-law noise model

In this paper we investigate the influence of a power-law noise model, also called noise, on the performance of a feed-forward neural network used to predict time series. We introduce an optimization procedure that optimizes the parameters the neural networks by maximizing the likelihood function based on the power-law model. We show that our optimization procedure minimizes the mean squared leading to an optimal prediction. Further, we present numerical results applying method to time series from the logistic map and the annual number of sunspots demonstrate that a power-law noise model gives better results than a Gaussian model.

Single mating in orchid bees (Euglossa, Apinae): implications for mate choice and social evolution

Neotropical orchid bees (Euglossini) are conspicuously different from other corbiculate bees (Apinae) in their lack of advanced sociality and in male use of acquired odors (fragrances) as pheromone-analogues. In both contexts, orchid bee mating systems, in particular the number of males a female mates with, are of great interest but are currently unknown. To assess female mating frequency in the genus Euglossa, we obtained nests from three species in Mexico and Panama and genotyped mothers and their brood at microsatellite DNA loci. In 26 out of 29 nests, genotypes of female brood were fully consistent with being descended from a singly mated mother. In nests with more than one adult female present, those adult females were frequently related, with genotypes being consistent with full sister-sister (r = 0.75) or mother-daughter (r = 0.5) relationships. Thus, our genetic data support the notions of female philopatry and nest-reuse in the genus Euglossa. Theoretically, single mating should promote the evolution of eusociality by maximizing the relatedness among individuals in a nest. However, in Euglossini this genetic incentive has not led to the formation of eusocial colonies as in other corbiculate bees, presumably due to differing ecological or physiological selective regimes. Finally, monandry in orchid bees is in agreement with the theory that females select a single best mate based on the male fragrance phenotype, which may contain information on male age, cognitive ability, and competitive strength.

Predicting community responses to perturbations in the face of imperfect knowledge and network complexity

How best to predict the effects of perturbations to ecological communities has been a long-standing goal for both applied and basic ecology. This quest has recently been revived by new empirical data, new analysis methods, and increased computing speed, with the promise that ecologically important insights may be obtainable from a limited knowledge of community interactions. We use empirically based and simulated networks of varying size and connectance to assess two limitations to predicting perturbation responses in multispecies communities: (1) the inaccuracy by which species interaction strengths are empirically quantified and (2) the indeterminacy of species responses due to indirect effects associated with network size and structure. We find that even modest levels of species richness and connectance (similar to 25 pairwise interactions) impose high requirements for interaction strength estimates because system indeterminacy rapidly overwhelms predictive insights. Nevertheless, even poorly estimated interaction strengths provide greater average predictive certainty than an approach that uses only the sign of each interaction. Our simulations provide guidance in dealing with the trade-offs involved in maximizing the utility of network approaches for predicting dynamics in multispecies communities.

It's all about relationships:A qualitative study of health researchers' perspectives of conducting interdisciplinary health research

Background. Interdisciplinary research has been promoted as an optimal research paradigm in the health sciences, yet little is known about how researchers experience interdisciplinarity in practice. This study sought to determine how interdisciplinary research was conceptualized and operationalized from the researcher's perspective and to better understand how best to facilitate interdisciplinary research success. Methods. Key informant interviews were conducted with health researchers with expertise or experience in conducting interdisciplinary research. Interviews were completed either in person or over the telephone using a semi-structured interview guide. Data collection occurred simultaneously with data analysis so that emerging themes could be explored in subsequent interviews. A content analysis approach was used. Results. Nineteen researchers took part in this study. Interdisciplinary research was conceptualized disparately between participants, and there was modest attention towards operationalization of interdisciplinary research. There was one overriding theme, "It's all about relationships", that emerged from the data. Within this theme, there were four related subthemes: 1) Involvement in interdisciplinary research; 2) Why do I do interdisciplinary research?; 3) Managing and fostering interdisciplinary relationships; and 4) The prickly side to interdisciplinary research. Together, these themes suggest that the choice to conduct interdisciplinary research, though often driven by the research question, is highly influenced by interpersonal and relationship-related factors. In addition, researchers preferred to engage in interdisciplinary research with those that they had already established relationships and where their role in the research process was clearly articulated. A focus on relationship building was seen as a strong facilitator of interdisciplinary success. Conclusion. Many health researchers experienced mixed reactions towards their involvement in interdisciplinary research. A well thought-out rationale for interdisciplinary research, and strategies to utilize the contribution of each researcher involved were seen as facilitators towards maximizing the benefits that could be derived from interdisciplinary research. © 2008 Nair et al; licensee BioMed Central Ltd.

Impact of composite manufacturing constraints on aerospace stiffened panel design

This paper reports on a design study assessing the impact of laminate manufacturing constraints on the structural performance and weight of composite stiffened panels. The study demonstrates that maximizing ply continuity results in weight penalties, while various geometric constraints related to manufacture and repair can be accommodated without significant weight penalties, potentially generating robust flexible designs.

How Much Multiuser Diversity is Required for Energy Limited Multiuser Systems?

Multiuser diversity (MUDiv) is one of the central concepts in multiuser (MU) systems. In particular, MUDiv allows for scheduling among users in order to eliminate the negative effects of unfavorable channel fading conditions of some users on the system performance. Scheduling, however, consumes energy (e.g., for making users' channel state information available to the scheduler). This extra usage of energy, which could potentially be used for data transmission, can be very wasteful, especially if the number of users is large. In this paper, we answer the question of how much MUDiv is required for energy limited MU systems. Focusing on uplink MU wireless systems, we develop MU scheduling algorithms which aim at maximizing the MUDiv gain. Toward this end, we introduce a new realistic energy model which accounts for scheduling energy and describes the distribution of the total energy between scheduling and data transmission stages. Using the fact that such energy distribution can be controlled by varying the number of active users, we optimize this number by either i) minimizing the overall system bit error rate (BER) for a fixed total energy of all users in the system or ii) minimizing the total energy of all users for fixed BER requirements. We find that for a fixed number of available users, the achievable MUDiv gain can be improved by activating only a subset of users. Using asymptotic analysis and numerical simulations, we show that our approach benefits from MUDiv gains higher than that achievable by generic greedy access algorithm, which is the optimal scheduling method for energy unlimited systems. © 2010 IEEE.

Power Allocation Strategies across N Orthogonal Channels at Both Source and Relay

A wireless relay network with one source, one relay and one destination is considered, where nodes communicate via N orthogonal channels. We develop optimal power allocation strategies at both the source and relay for maximizing the overall source-destination capacity under individual power constraints at the source and relay. Some properties of the optimal solution are studied. © 2012 IEEE.

Orthogonal space-time block coded rate-adaptive modulation with outdated feedbasck

Abstract-Channel state information (CSI) at the transmitter can be used to adapt transmission rate or antenna gains in multi-antenna systems. We propose a rate-adaptive M-QAM scheme equipped with orthogonal space-time block coding with simple outdated, finite-rate feedback over independent flat fading channels. We obtain closed-form expressions for the average BER and throughput for our scheme, and analyze the effects of possibly delayed feedback on the performance gains. We derive optimal switching thresholds maximizing the average throughput under average and outage BER constraints with outdated feedback. Our numerical results illustrate the immunity of our optimal thresholds to delayed feedback.

Intermediate Care: The Role of Medicines Management

Healthcare systems worldwide are facing an unprecedented demographic change as globally, the number of older people will triple to 2 billion by the year 2050. The resulting pressures on acute services have been instrumental in the development of intermediate care (IC) as a new healthcare model, which has its origins in the National Health Service in the UK. IC is an umbrella term for patient services that do not require the resources of a general hospital but are beyond the scope of a traditional primary care team. IC aims to promote timely discharge from hospital, prevent unnecessary hospital admissions and reduce the need for long-term residential care by optimizing functional independence. Various healthcare providers around the world have adopted similar models of care to manage changing healthcare needs. Polypharmacy, along with age-related changes, places older people at an increased risk of adverse drug events, including inappropriate prescribing, which has been shown to be prevalent in this population in other healthcare settings. Medicines management (the practice of maximizing health through optimal use of medicines) of older people has been discussed in the literature in a variety of settings; however, its place within IC is largely unknown. Despite IC being a multidisciplinary healthcare model, there is a lack of evidence to suggest that enhanced pharmaceutical involvement is core to the service provided within IC. This review article highlights the gap in the literature surrounding medicines management within IC and identifies potential solutions aimed at improving patient outcomes in this setting.

A multi-output two-stage locally regularized model construction method using the extreme learning machine

This paper investigates the construction of linear-in-the-parameters (LITP) models for multi-output regression problems. Most existing stepwise forward algorithms choose the regressor terms one by one, each time maximizing the model error reduction ratio. The drawback is that such procedures cannot guarantee a sparse model, especially under highly noisy learning conditions. The main objective of this paper is to improve the sparsity and generalization capability of a model for multi-output regression problems, while reducing the computational complexity. This is achieved by proposing a novel multi-output two-stage locally regularized model construction (MTLRMC) method using the extreme learning machine (ELM). In this new algorithm, the nonlinear parameters in each term, such as the width of the Gaussian function and the power of a polynomial term, are firstly determined by the ELM. An initial multi-output LITP model is then generated according to the termination criteria in the first stage. The significance of each selected regressor is checked and the insignificant ones are replaced at the second stage. The proposed method can produce an optimized compact model by using the regularized parameters. Further, to reduce the computational complexity, a proper regression context is used to allow fast implementation of the proposed method. Simulation results confirm the effectiveness of the proposed technique. © 2013 Elsevier B.V.

Secure Multiuser Multiple Amplify-and-Forward Relay Networks in Presence of Multiple Eavesdroppers

In this paper, we study the information-theoretical security of a downlink multiuser cooperative relaying network with multiple intermediate amplify-and-forward (AF) relays, where there exist multiple eavesdroppers which can overhear the message. To prevent the wiretap and strength the network security, we select one best relay and user pair, so that the selected user can receive the message from the base station assisted by the selected relay. The relay and user selection is performed by maximizing the ratio of the received signal-to-noise ratio (SNR) at the user to the eavesdroppers, which is based on both the main and eavesdropper links. For the considered system, we derive the closed-form expression of the secrecy outage probability, and provide the asymptotic expression in high main-to-eavesdropper ratio (MER) region. From the asymptotic analysis, we can find that the system diversity order is equivalent to the number of relays regardless of the number of users and eavesdroppers.

Secure Multiuser Communications in Multiple Amplify-and-Forward Relay Networks

This paper proposes relay selection in order to increase the physical layer security in multiuser cooperative relay networks with multiple amplify-and-forward (AF) relays, in the presence of multiple eavesdroppers. To strengthen the network security against eavesdropping attack, we present three criteria to select the best relay and user pair. Specifically, criterion I and II study the received signal-to-noise ratio (SNR) at the receivers, and perform the selection by maximizing the SNR ratio of the user to the eavesdroppers. To this end, criterion I relies on both the main and eavesdropper links, while criterion II relies on the main links only. Criterion III is the standard max-min selection criterion,
which maximizes the minimum of the dual-hop channel gains of main links. For the three selection criteria, we examine the system secrecy performance by deriving the analytical expressions for the secrecy outage probability. We also derive the asymptotic analysis for the secrecy outage probability with high main-to eavesdropper ratio (MER). From the asymptotic analysis, an interesting observation is reached: for each criterion, the system diversity order is equivalent to the number of relays regardless of the number of users and eavesdroppers.

Efficient Marginal Likelihood Computation for Gaussian Process Regression

In a Bayesian learning setting, the posterior distribution of a predictive model arises from a trade-off between its prior distribution and the conditional likelihood of observed data. Such distribution functions usually rely on additional hyperparameters which need to be tuned in order to achieve optimum predictive performance; this operation can be efficiently performed in an Empirical Bayes fashion by maximizing the posterior marginal likelihood of the observed data. Since the score function of this optimization problem is in general characterized by the presence of local optima, it is necessary to resort to global optimization strategies, which require a large number of function evaluations. Given that the evaluation is usually computationally intensive and badly scaled with respect to the dataset size, the maximum number of observations that can be treated simultaneously is quite limited. In this paper, we consider the case of hyperparameter tuning in Gaussian process regression. A straightforward implementation of the posterior log-likelihood for this model requires O(N^3) operations for every iteration of the optimization procedure, where N is the number of examples in the input dataset. We derive a novel set of identities that allow, after an initial overhead of O(N^3), the evaluation of the score function, as well as the Jacobian and Hessian matrices, in O(N) operations. We prove how the proposed identities, that follow from the eigendecomposition of the kernel matrix, yield a reduction of several orders of magnitude in the computation time for the hyperparameter optimization problem. Notably, the proposed solution provides computational advantages even with respect to state of the art approximations that rely on sparse kernel matrices.

Monte Carlo calculations of quantum yield in inhomogeneous PtSi/p-Si Schottky barriers

Monte Carlo calculations of quantum yield in PtSi/p-Si infrared detectors are carried out taking into account the presence of a spatially distributed barrier potential. In the 1-4 mu m wavelength range it is found that the spatial inhomogeneity of the barrier has no significant effect on the overall device photoresponse. However, above lambda = 4.0 mu m and particularly as the cut-off wavelength (lambda approximate to 5.5 mu m) is approached, these calculations reveal a difference between the homogeneous and inhomogeneous barrier photoresponse which becomes increasingly significant and exceeds 50% at lambda = 5.3 mu m. It is, in fact, the inhomogeneous barrier which displays an increased photoyield, a feature that is confirmed by approximate analytical calculations assuming a symmetric Gaussian spatial distribution of the barrier. Furthermore, the importance of the silicide layer thickness in optimizing device efficiency is underlined as a trade-off between maximizing light absorption in the silicide layer and optimizing the internal yield. The results presented here address important features which determine the photoyield of PtSi/Si Schottky diodes at energies below the Si absorption edge and just above the Schottky barrier height in particular.

Identifying PV module mismatch faults by a thermography-based temperature distribution analysis

Photovoltaic (PV) solar power generation is proven to be effective and sustainable but is currently hampered by relatively high costs and low conversion efficiency. This paper addresses both issues by presenting a low-cost and efficient temperature distribution analysis for identifying PV module mismatch faults by thermography. Mismatch faults reduce the power output and cause potential damage to PV cells. This paper first defines three fault categories in terms of fault levels, which lead to different terminal characteristics of the PV modules. The investigation of three faults is also conducted analytically and experimentally, and maintenance suggestions are also provided for different fault types. The proposed methodology is developed to combine the electrical and thermal characteristics of PV cells subjected to different fault mechanisms through simulation and experimental tests. Furthermore, the fault diagnosis method can be incorporated into the maximum power point tracking schemes to shift the operating point of the PV string. The developed technology has improved over the existing ones in locating the faulty cell by a thermal camera, providing a remedial measure, and maximizing the power output under faulty conditions.