Adaptive filtering for high quality HMM based speech synthesis

In this work an adaptive filtering scheme based on a dual Discrete Kalman Filtering (DKF) is proposed for Hidden Markov Model (HMM) based speech synthesis quality enhancement. The objective is to improve signal smoothness across HMMs and their related states and to reduce artifacts due to acoustic model's limitations. Both speech and artifacts are modelled by an autoregressive structure which provides an underlying time frame dependency and improves time-frequency resolution. Themodel parameters are arranged to obtain a combined state-space model and are also used to calculate instantaneous power spectral density estimates. The quality enhancement is performed by a dual discrete Kalman filter that simultaneously gives estimates for the models and the signals. The system's performance has been evaluated using mean opinion score tests and the proposed technique has led to improved results.

The fitting of the German yield curve: A dynamic approach using latent, macroeconomic and stock market variables

A Work Project, presented as part of the requirements for the Award of a Masters Degree in Finance from the NOVA – School of Business and Economics

Shocks, Monetary Policy and Institutions: Explaining Unemployment Persistence in Persistence in “Europe” and the United States

This paper examines the rise in European unemployment since the 1970s by introducing endogenous growth into an otherwise standard New Keynesian model with capital accumulation and unemployment. We subject the model to an uncorrelated cost push shock, in order to mimic a scenario akin to the one faced by central banks at the end of the 1970s. Monetary policy implements a disinfl ation by following an interest feedback rule calibrated to an estimate of a Bundesbank reaction function. 40 quarters after the shock has vanished, unemployment is still about 1.8 percentage points above its steady state. Our model also broadly reproduces cross country differences in unemployment by drawing on cross country differences in the size of cost push shock and the associated disinfl ation, the monetary policy reaction function and the wage setting structure.

UK Macroeconomic Forecasting with Many Predictors: Which Models Forecast Best and When Do They Do So?

Block factor methods offer an attractive approach to forecasting with many predictors. These extract the information in these predictors into factors reflecting different blocks of variables (e.g. a price block, a housing block, a financial block, etc.). However, a forecasting model which simply includes all blocks as predictors risks being over-parameterized. Thus, it is desirable to use a methodology which allows for different parsimonious forecasting models to hold at different points in time. In this paper, we use dynamic model averaging and dynamic model selection to achieve this goal. These methods automatically alter the weights attached to different forecasting models as evidence comes in about which has forecast well in the recent past. In an empirical study involving forecasting output growth and inflation using 139 UK monthly time series variables, we find that the set of predictors changes substantially over time. Furthermore, our results show that dynamic model averaging and model selection can greatly improve forecast performance relative to traditional forecasting methods.

UK Macroeconomic Forecasting with Many Predictors: Which Models Forecast Best and When Do They Do So?

Block factor methods offer an attractive approach to forecasting with many predictors. These extract the information in these predictors into factors reflecting different blocks of variables (e.g. a price block, a housing block, a financial block, etc.). However, a forecasting model which simply includes all blocks as predictors risks being over-parameterized. Thus, it is desirable to use a methodology which allows for different parsimonious forecasting models to hold at different points in time. In this paper, we use dynamic model averaging and dynamic model selection to achieve this goal. These methods automatically alter the weights attached to different forecasting model as evidence comes in about which has forecast well in the recent past. In an empirical study involving forecasting output and inflation using 139 UK monthly time series variables, we find that the set of predictors changes substantially over time. Furthermore, our results show that dynamic model averaging and model selection can greatly improve forecast performance relative to traditional forecasting methods.

Large Time-Varying Parameter VARs

In this paper we develop methods for estimation and forecasting in large timevarying parameter vector autoregressive models (TVP-VARs). To overcome computational constraints with likelihood-based estimation of large systems, we rely on Kalman filter estimation with forgetting factors. We also draw on ideas from the dynamic model averaging literature and extend the TVP-VAR so that its dimension can change over time. A final extension lies in the development of a new method for estimating, in a time-varying manner, the parameter(s) of the shrinkage priors commonly-used with large VARs. These extensions are operationalized through the use of forgetting factor methods and are, thus, computationally simple. An empirical application involving forecasting inflation, real output, and interest rates demonstrates the feasibility and usefulness of our approach.

Using VARs and TVP-VARs with Many Macroeconomic Variables

This paper discusses the challenges faced by the empirical macroeconomist and methods for surmounting them. These challenges arise due to the fact that macroeconometric models potentially include a large number of variables and allow for time variation in parameters. These considerations lead to models which have a large number of parameters to estimate relative to the number of observations. A wide range of approaches are surveyed which aim to overcome the resulting problems. We stress the related themes of prior shrinkage, model averaging and model selection. Subsequently, we consider a particular modelling approach in detail. This involves the use of dynamic model selection methods with large TVP-VARs. A forecasting exercise involving a large US macroeconomic data set illustrates the practicality and empirical success of our approach.

Disputes, Debt and Equity

We show how the prospect of disputes over firms’ revenue reports promotes debt financing over equity. These findings are presented within a costly state verification model with a risk averse entrepreneur. The prospect of disputes encourages incentive regimes which limit penalties and avoid stochastic monitoring, even when the lender can commit to stochastic enforcement strategies. Consequently, optimal contracts shift away from equity and toward standard debt. For a useful special case of the model, closed form solutions are presented for leverage and consumption allocations under efficient debt contracts.

Cell-based computational modeling of vascular morphogenesis using Tissue Simulation Toolkit.

Computational modeling has become a widely used tool for unraveling the mechanisms of higher level cooperative cell behavior during vascular morphogenesis. However, experimenting with published simulation models or adding new assumptions to those models can be daunting for novice and even for experienced computational scientists. Here, we present a step-by-step, practical tutorial for building cell-based simulations of vascular morphogenesis using the Tissue Simulation Toolkit (TST). The TST is a freely available, open-source C++ library for developing simulations with the two-dimensional cellular Potts model, a stochastic, agent-based framework to simulate collective cell behavior. We will show the basic use of the TST to simulate and experiment with published simulations of vascular network formation. Then, we will present step-by-step instructions and explanations for building a recent simulation model of tumor angiogenesis. Demonstrated mechanisms include cell-cell adhesion, chemotaxis, cell elongation, haptotaxis, and haptokinesis.

Electronic couplings and on-site energies for hole transfer in DNA: systematic quantum mechanical/molecular dynamic study

The electron hole transfer (HT) properties of DNA are substantially affected by thermal fluctuations of the π stack structure. Depending on the mutual position of neighboring nucleobases, electronic coupling V may change by several orders of magnitude. In the present paper, we report the results of systematic QM/molecular dynamic (MD) calculations of the electronic couplings and on-site energies for the hole transfer. Based on 15 ns MD trajectories for several DNA oligomers, we calculate the average coupling squares 〈 V2 〉 and the energies of basepair triplets X G+ Y and X A+ Y, where X, Y=G, A, T, and C. For each of the 32 systems, 15 000 conformations separated by 1 ps are considered. The three-state generalized Mulliken-Hush method is used to derive electronic couplings for HT between neighboring basepairs. The adiabatic energies and dipole moment matrix elements are computed within the INDO/S method. We compare the rms values of V with the couplings estimated for the idealized B -DNA structure and show that in several important cases the couplings calculated for the idealized B -DNA structure are considerably underestimated. The rms values for intrastrand couplings G-G, A-A, G-A, and A-G are found to be similar, ∼0.07 eV, while the interstrand couplings are quite different. The energies of hole states G+ and A+ in the stack depend on the nature of the neighboring pairs. The X G+ Y are by 0.5 eV more stable than X A+ Y. The thermal fluctuations of the DNA structure facilitate the HT process from guanine to adenine. The tabulated couplings and on-site energies can be used as reference parameters in theoretical and computational studies of HT processes in DNA

Cost-effectiveness of low-molecular-weight heparin for treatment of pulmonary embolism.

BACKGROUND: Low-molecular-weight heparin (LMWH) appears to be safe and effective for treating pulmonary embolism (PE), but its cost-effectiveness has not been assessed. METHODS: We built a Markov state-transition model to evaluate the medical and economic outcomes of a 6-day course with fixed-dose LMWH or adjusted-dose unfractionated heparin (UFH) in a hypothetical cohort of 60-year-old patients with acute submassive PE. Probabilities for clinical outcomes were obtained from a meta-analysis of clinical trials. Cost estimates were derived from Medicare reimbursement data and other sources. The base-case analysis used an inpatient setting, whereas secondary analyses examined early discharge and outpatient treatment with LMWH. Using a societal perspective, strategies were compared based on lifetime costs, quality-adjusted life-years (QALYs), and the incremental cost-effectiveness ratio. RESULTS: Inpatient treatment costs were higher for LMWH treatment than for UFH (dollar 13,001 vs dollar 12,780), but LMWH yielded a greater number of QALYs than did UFH (7.677 QALYs vs 7.493 QALYs). The incremental costs of dollar 221 and the corresponding incremental effectiveness of 0.184 QALYs resulted in an incremental cost-effectiveness ratio of dollar 1,209/QALY. Our results were highly robust in sensitivity analyses. LMWH became cost-saving if the daily pharmacy costs for LMWH were < dollar 51, if > or = 8% of patients were eligible for early discharge, or if > or = 5% of patients could be treated entirely as outpatients. CONCLUSION: For inpatient treatment of PE, the use of LMWH is cost-effective compared to UFH. Early discharge or outpatient treatment in suitable patients with PE would lead to substantial cost savings.

A low dimensional Kalman filter for systems with lagged observables

This note describes how the Kalman filter can be modified to allow for thevector of observables to be a function of lagged variables without increasing the dimensionof the state vector in the filter. This is useful in applications where it is desirable to keepthe dimension of the state vector low. The modified filter and accompanying code (whichnests the standard filter) can be used to compute (i) the steady state Kalman filter (ii) thelog likelihood of a parameterized state space model conditional on a history of observables(iii) a smoothed estimate of latent state variables and (iv) a draw from the distribution oflatent states conditional on a history of observables.

Selecting statistical models to study the relationship between soybean yield and soil physical properties

Statistical models allow the representation of data sets and the estimation and/or prediction of the behavior of a given variable through its interaction with the other variables involved in a phenomenon. Among other different statistical models, are the autoregressive state-space models (ARSS) and the linear regression models (LR), which allow the quantification of the relationships among soil-plant-atmosphere system variables. To compare the quality of the ARSS and LR models for the modeling of the relationships between soybean yield and soil physical properties, Akaike's Information Criterion, which provides a coefficient for the selection of the best model, was used in this study. The data sets were sampled in a Rhodic Acrudox soil, along a spatial transect with 84 points spaced 3 m apart. At each sampling point, soybean samples were collected for yield quantification. At the same site, soil penetration resistance was also measured and soil samples were collected to measure soil bulk density in the 0-0.10 m and 0.10-0.20 m layers. Results showed autocorrelation and a cross correlation structure of soybean yield and soil penetration resistance data. Soil bulk density data, however, were only autocorrelated in the 0-0.10 m layer and not cross correlated with soybean yield. The results showed the higher efficiency of the autoregressive space-state models in relation to the equivalent simple and multiple linear regression models using Akaike's Information Criterion. The resulting values were comparatively lower than the values obtained by the regression models, for all combinations of explanatory variables.

Bi-planar 2D-to-3D registration in Fourier domain for stereoscopic X-ray motion tracking

In this paper we present a new method to track bonemovements in stereoscopic X-ray image series of the kneejoint. The method is based on two different X-ray imagesets: a rotational series of acquisitions of the stillsubject knee that will allow the tomographicreconstruction of the three-dimensional volume (model),and a stereoscopic image series of orthogonal projectionsas the subject performs movements. Tracking the movementsof bones throughout the stereoscopic image series meansto determine, for each frame, the best pose of everymoving element (bone) previously identified in the 3Dreconstructed model. The quality of a pose is reflectedin the similarity between its simulated projections andthe actual radiographs. We use direct Fourierreconstruction to approximate the three-dimensionalvolume of the knee joint. Then, to avoid the expensivecomputation of digitally rendered radiographs (DRR) forpose recovery, we reformulate the tracking problem in theFourier domain. Under the hypothesis of parallel X-raybeams, we use the central-slice-projection theorem toreplace the heavy 2D-to-3D registration of projections inthe signal domain by efficient slice-to-volumeregistration in the Fourier domain. Focusing onrotational movements, the translation-relevant phaseinformation can be discarded and we only consider scalarFourier amplitudes. The core of our motion trackingalgorithm can be implemented as a classical frame-wiseslice-to-volume registration task. Preliminary results onboth synthetic and real images confirm the validity ofour approach.

Multiway Array Decomposition Analysis of EEGs in Alzheimer’s Disease

Methods for the extraction of features from physiological datasets are growing needs as clinical investigations of Alzheimer’s disease (AD) in large and heterogeneous population increase. General tools allowing diagnostic regardless of recording sites, such as different hospitals, are essential and if combined to inexpensive non-invasive methods could critically improve mass screening of subjects with AD. In this study, we applied three state of the art multiway array decomposition (MAD) methods to extract features from electroencephalograms (EEGs) of AD patients obtained from multiple sites. In comparison to MAD, spectral-spatial average filter (SSFs) of control and AD subjects were used as well as a common blind source separation method, algorithm for multiple unknown signal extraction (AMUSE). We trained a feed-forward multilayer perceptron (MLP) to validate and optimize AD classification from two independent databases. Using a third EEG dataset, we demonstrated that features extracted from MAD outperformed features obtained from SSFs AMUSE in terms of root mean squared error (RMSE) and reaching up to 100% of accuracy in test condition. We propose that MAD maybe a useful tool to extract features for AD diagnosis offering great generalization across multi-site databases and opening doors to the discovery of new characterization of the disease.