In this letter, we propose a novel method of refining the time-domain synthesis individual source estimates from single channel mixture. Employing closed-loop architecture, algorithm refines each by iteratively estimating phase sources, given magnitude spectra and The performance is evaluated for harmonic musical mixtures, considerable improvements to synthesized are obtained relative binary masking, accurate spectra.

Accurate predictions of surface waves and subsequent wave-induced vessel motion have the potential to improve safety efficiency for a wide range offshore operations, such as active control wave energy converters floating wind turbines. In this paper, an Auto-Regressive model is proposed predict motions based on measured time sequences at particular location. Based numerically synthesized data, it shown that band-pass filtering with single cut-off frequency can accuracy predictions. These...

Surface wave predictions for several periods in advance are crucial optimizing a wide range of offshore applications. This work focuses on the potential application active control energy converters (WECs), which can dramatically enhance efficiency power generation. Field tests were conducted Southern Ocean Albany, Western Australia. We compared two prediction models: physics-based algebraic model and machine learning-based Artificial Neural Network (ANN) model. Although standard ANN is found...

This paper introduces a novel approach for acoustic scene analysis by exploiting an ensemble of statistics extracted from sub-band domain multi-hypothesis echo canceler (SDMH-AEC). A well-designed SDMH-AEC employs multiple adaptive filtering strategies with potentially complementary behaviours during convergence, perturbations, and steady-state conditions. By aggregating across the sub-bands, we derive feature vector that exhibits strong discriminative power distinguishing different events...

Evidence accumulation models (EAMs) are an important class of cognitive used to analyze both response time and choice data recorded from decision-making tasks. Developments in estimation procedures have helped EAMs become basic scientific applications solution-focused applied work. Hierarchical Bayesian frameworks for the linear ballistic accumulator (LBA) model diffusion decision (DDM) been widely used, but still suffer some key limitations, particularly large sample sizes, with many...

Many psychological experiments have subjects repeat a task to gain the statistical precision required test quantitative theories of performance. In such experiments, time-on-task can sizable effects on performance, changing processes under investigation. Most research has either ignored these changes, treating underlying process as static, or sacrificed some content models for simplicity. We use particle Markov chain Monte-Carlo methods study psychologically plausible time-varying changes in...

Accurate surface wave predictions have the potential to greatly enhance safety and efficiency of many offshore applications, such as active control energy converters floating wind turbines. However, real-time prediction becomes increasingly challenging when large directional spreading is considered. To address this challenge, present study introduces a machine learning model that utilizes an Artificial Neural Network (ANN) for predicting moderate waves. Linear, short-crested time histories...

Model comparison is the cornerstone of theoretical progress in psychological research. Common practice overwhelmingly relies on tools that evaluate competing models by balancing in-sample descriptive adequacy against model flexibility, with modern approaches advocating use marginal likelihood for hierarchical cognitive models. Cross-validation another popular approach but its implementation remains out reach evaluated a Bayesian framework, major hurdle being prohibitive computational cost....

Abstract This paper proposes a novel link between credit markets and uncertainty shocks. We introduce role for via collateral constraints in an otherwise standard real business cycle (RBC) model show that increase triggers precautionary response interacts with the constraint to generate simultaneous decline output, consumption, investment, wages, hours; feature previous work on shocks without is unable produce flexible‐price environment. also empirically test theoretical predictions...

This article studies the use of asymmetric loss functions for optimal prediction positive-valued spatial processes. We focus on family power-divergence with properties such as continuity, convexity, connection to well known divergence measures, and ability control asymmetry behaviour function via a power parameter. The functions, (OPD) predictors, related measures uncertainty quantification are studied. In addition, we examine in general notion defined processes define an measure, which...

Summary Data from large surveys are often supplemented with sampling weights that designed to reflect unequal probabilities of response and selection inherent in complex survey methods. We propose two methods for Bayesian estimation parametric models a setting where the data available, but information on how were constructed is unavailable. The first approach simply replace likelihood pseudo formulation Bayes theorem. This proven lead consistent estimator also leads credible intervals suffer...

This article addresses the problem of efficient Bayesian inference in dynamic systems using particle methods and makes a number contributions. First, we develop correlated pseudo-marginal (CPM) approach for state space (SS) models that is based on filtering disturbances, rather than states. useful when transition density intractable or inefficient to compute, also dimension disturbance lower state. Second, propose block (BPM) method uses as estimate likelihood average G independent unbiased...

This article extends the literature on copulas with discrete or continuous marginals to case where some of are a mixture and components. We do so by carefully defining likelihood as density observations respect mixed measure. The treatment is quite general, although we focus mixtures Gaussian Archimedean copulas. inference Bayesian estimation carried out Markov chain Monte Carlo. illustrate methodology algorithms applying them estimate multivariate income dynamics model. Supplementary...

Efficient estimation methods for simultaneous autoregressive (SAR) models with missing data in the response variable have been well explored literature. A common practice is introducing measurement error into SAR to separate noise component from spatial process. However, prior studies not considered incorporating data. Maximum likelihood such models, especially large datasets, poses significant computational challenges. This paper proposes an efficient likelihood-based method, marginal...

It is well known that the spectral envelope a perceptually salient attribute in musical instrument timbre perception. While number of studies have explored discrimination thresholds for changes to envelope, question how sensitivity varies as function center frequency and bandwidth instruments has yet be addressed. In this paper two-alternative forced-choice experiment was conducted observe perceptual modifications made on trumpet, clarinet viola sounds. The involved attenuating 14 bands each...

The stochastic volatility (SV) model and its variants are widely used in the financial sector, while recurrent neural network (RNN) models successfully many large-scale industrial applications of deep learning. We combine these two methods a nontrivial way propose model, which we call statistical (SR-SV) to capture dynamics volatility. proposed is able complex effects, for example, nonlinearity long-memory auto-dependence, overlooked by conventional SV models, statistically interpretable has...