Additive Error Term Research Articles

Semi-Supervised Triply Robust Inductive Transfer Learning

In this work, we propose a Semi-supervised Triply Robust Inductive transFer LEarning (STRIFLE) approach, which integrates heterogeneous data from a label-rich source population and a label-scarce target population and uses a large amount of unlabeled data simultaneously to improve the learning accuracy in the target population. Specifically, we consider a high dimensional covariate shift setting and employ two nuisance models, a density ratio model and an imputation model, to combine transfer learning and surrogate-assisted semi-supervised learning strategies effectively and achieve triple robustness. While the STRIFLE approach assumes the target and source populations to share the same conditional distribution of outcome Y given both the surrogate features S and predictors X , it allows the true underlying model of Y⏧ X to differ between the two populations due to the potential covariate shift in S and X . Different from double robustness, even if both nuisance models are misspecified or the distribution of Y⏧ S , X is not the same between the two populations when the shifted source population and the target population share enough similarities, the triply robust STRIFLE estimator can still partially use the source population when the shifted source population and the target population share enough similarities. Moreover, it is guaranteed to be no worse than the target-only surrogate-assisted semi-supervised estimator with an additional error term from transferability detection. These desirable properties of our estimator are established theoretically and verified in finite samples via extensive simulation studies. We use the STRIFLE estimator to train a Type II diabetes polygenic risk prediction model for the African American target population by transferring knowledge from electronic health records linked genomic data observed in a larger European source population. Supplementary materials for this article are available online, including a standardized description of the materials available for reproducing the work.

Stress-related topology optimization with multilevel smoothed isogeometric densities and Bézier elements

Isogeometric analysis (IGA) provides a feasible technique to seamlessly integrate computer-aided design (CAD) into the existing finite element analysis. In this article, a Bézier elements-based IGA method is established to address the stress-related topology optimization of structures, which incorporates geometrical representation, structural analysis and topology optimization into a unified process as well as can naturally handle the curvilinear edge elements for the stress problems. To enhance the smoothness of the structural boundaries, a multilevel isogeometric density field (IDF) defined at control points (CPs) is designed by utilizing the values of the non-uniform rational B-splines (NURBS) basis functions from the previous design iteration. A global stress measure is approximated by a smooth maximum function with an additional error term. The k-means clustering originally from signal processing is used to partition the stress observations with dynamic centroids and indices, which improves the accuracy of the sensitivity results. The sensitivity analyses of the stress-related problems are derived from the multilevel IDF model. Three kinds of numerical results are provided to illustrate the effectiveness of the proposed method to solve stress-related problems via the multilevel IDF model for achieving the optimal solution with smooth boundaries.

Frequency-Modulated OFDM: A New Waveform for High-Mobility Wireless Communications

This paper introduces a novel waveform for wireless communication, denoted as Frequency-Modulated Orthogonal Frequency-Division Multiplexing (FM-OFDM). By frequency-modulating an OFDM signal and further defining a suitable cutoff subcarrier, the resulting constant-envelope waveform (ensuring a low Peak to Average Power Ratio) exhibits strong robustness not only to Doppler, but also to phase noise and carrier frequency offsets. These impairments introduce an additive error term at the information-bearing subcarriers that is mostly concentrated at the lowest part of the spectrum. No inter-carrier interference is thus present, and impairments can be easily overcome by mapping the information above a cutoff subcarrier. Theoretical expressions and numerical results prove the superiority of FM-OFDM over the state of the art, particularly in high-speed Rayleigh channels and/or high phase noise.

On the influence of the nonlinear term in the numerical approximation of Incompressible Flows by means of proper orthogonal decomposition methods

We consider proper orthogonal decomposition (POD) methods to approximate the incompressible Navier–Stokes equations. We study the case in which one discretization for the nonlinear term is used in the snapshots (that are computed with a full order method (FOM)) and a different discretization of the nonlinear term is applied in the POD method. We prove that an additional error term appears in this case, compared with the case in which the same discretization of the nonlinear term is applied for both the FOM and the POD methods. However, the added term has the same size as the error coming from the FOM so that the rate of convergence of the POD method is barely affected. We analyze the case in which we add grad–div stabilization to both the FOM and the POD methods because it allows to get error bounds with constants independent of inverse powers of the viscosity. We also study the case in which no stabilization is added. Some numerical experiments support the theoretical analysis.

Quantitatively characterizing reflexive responses to pitch perturbations.

Reflexive pitch perturbation experiments are commonly used to investigate the neural mechanisms underlying vocal motor control. In these experiments, the fundamental frequency-the acoustic correlate of pitch-of a speech signal is shifted unexpectedly and played back to the speaker via headphones in near real-time. In response to the shift, speakers increase or decrease their fundamental frequency in the direction opposing the shift so that their perceived pitch is closer to what they intended. The goal of the current work is to develop a quantitative model of responses to reflexive perturbations that can be interpreted in terms of the physiological mechanisms underlying the response and that captures both group-mean data and individual subject responses. A model framework was established that allowed the specification of several models based on Proportional-Integral-Derivative and State-Space/Directions Into Velocities of Articulators (DIVA) model classes. The performance of 19 models was compared in fitting experimental data from two published studies. The models were evaluated in terms of their ability to capture both population-level responses and individual differences in sensorimotor control processes. A three-parameter DIVA model performed best when fitting group-mean data from both studies; this model is equivalent to a single-rate state-space model and a first-order low pass filter model. The same model also provided stable estimates of parameters across samples from individual subject data and performed among the best models to differentiate between subjects. The three parameters correspond to gains in the auditory feedback controller's response to a perceived error, the delay of this response, and the gain of the somatosensory feedback controller's "resistance" to this correction. Excellent fits were also obtained from a four-parameter model with an additional auditory velocity error term; this model was better able to capture multi-component reflexive responses seen in some individual subjects. Our results demonstrate the stereotyped nature of an individual's responses to pitch perturbations. Further, we identified a model that captures population responses to pitch perturbations and characterizes individual differences in a stable manner with parameters that relate to underlying motor control capabilities. Future work will evaluate the model in characterizing responses from individuals with communication disorders.

Assessment of polymer feedback coupling approaches in simulation of viscoelastic fluids using the lattice Boltzmann method

The lattice Boltzmann method (LBM) has emerged as a popular approach to simulate complex non-Newtonian fluid mechanics problems, such as those involving viscoelastic fluids. In these cases, the LBM solver often requires modelling additional non-linear and non-ideal hydrodynamic contributions to the flow resulting from the addition of polymer molecules. These additional contributions, commonly referred to as the polymer feedback, are either incorporated directly in the Boltzmann stress (pressure method) or alternatively in the momentum field as a forcing contribution (forcing method). The choice of coupling approach has in previous works been strongly speculated to impact the numerical accuracy and stability of LBM solvers for viscoelastic fluids, however no formal comparison has been conducted. In this work, we compare the two popular polymer feedback coupling approaches, as well as the various schemes commonly used in the forcing method, namely, the explicit forcing (EF) scheme (He et al., 1998), the Guo et al. scheme (Guo et al., 2002), and the exact-difference-method (EDM) (Kupershtokh et al., 2009). First, through a theoretical comparison in recovering the macroscopic equations, we show that both the EF and Guo et al. forcing schemes retain the exact hydrodynamic representation up to second-order, whereas the EDM scheme includes additional lattice error terms. Similarly, the pressure method retains additional unphysical terms, which are known to violate Galilean invariance. Through numerical tests, involving the steady and time-dependent Poiseuille flow cases, as well as the four-roll mill problem under both steady and transient regimes, it is shown that the forcing method has enhanced numerical accuracy and stability at varying elastic effects. Alternatively, owing to the violation of Galilean invariance, the pressure method proved to be unstable even for moderate elastic effects. The obtained results are useful in determining the optimal polymer feedback coupling approach when simulating viscoelastic fluids using LBM, as well as further testing the generality of previous LBM findings on non-ideal coupling approaches for viscoelastic fluids.

Online Apprenticeship Learning

In Apprenticeship Learning (AL), we are given a Markov Decision Process (MDP) without access to the cost function. Instead, we observe trajectories sampled by an expert that acts according to some policy. The goal is to find a policy that matches the expert's performance on some predefined set of cost functions. We introduce an online variant of AL (Online Apprenticeship Learning; OAL), where the agent is expected to perform comparably to the expert while interacting with the environment. We show that the OAL problem can be effectively solved by combining two mirror descent based no-regret algorithms: one for policy optimization and another for learning the worst case cost. By employing optimistic exploration, we derive a convergent algorithm with O(sqrt(K)) regret, where K is the number of interactions with the MDP, and an additional linear error term that depends on the amount of expert trajectories available. Importantly, our algorithm avoids the need to solve an MDP at each iteration, making it more practical compared to prior AL methods. Finally, we implement a deep variant of our algorithm which shares some similarities to GAIL, but where the discriminator is replaced with the costs learned by OAL. Our simulations suggest that OAL performs well in high dimensional control problems.

Finite-Sample Analysis for Decentralized Batch Multiagent Reinforcement Learning With Networked Agents

Despite the increasing interest in multiagent reinforcement learning (MARL) in multiple communities, understanding its theoretical foundation has long been recognized as a challenging problem. In this article, we address this problem by providing a finite-sample analysis for decentralized batch MARL. Specifically, we consider a type of mixed MARL setting with both cooperative and competitive agents, where two teams of agents compete in a zero-sum game setting, while the agents within each team collaborate by communicating over a time-varying network. This setting covers many conventional MARL settings in the literature. We then develop batch MARL algorithms that can be implemented in a decentralized fashion, and quantify the finite-sample errors of the estimated action-value functions. Our error analysis captures how the function class, the number of samples within each iteration, and the number of iterations determine the statistical accuracy of the proposed algorithms. Our results, compared to the finite-sample bounds for single-agent reinforcement learning, involve additional error terms caused by decentralized computation, which is inherent in our decentralized MARL setting. This article provides the first finite-sample analysis for batch MARL, a step toward rigorous theoretical understanding of general MARL algorithms in the finite-sample regime.

A simple test of completeness in a class of nonparametric specification

This paper provides a test for completeness in a class of nonparametric specification with an additive and independent error term. It is known that such a nonparametric location family of functions is complete if and only if the characteristic function of the error term has no zeros on the real line. Because a zero of the error characteristic function implies that of an observed marginal distribution, we propose a simple test for zeros of characteristic function of the observed distribution, in which rejection of the null hypothesis implies the completeness. This test is applicable to many popular settings, such as nonparametric regression models with instrumental variables, and nonclassical measurement error models. We describe the asymptotic behavior of the tests under the null and alternative hypotheses and investigate the finite sample properties of the proposed test through a Monte Carlo study. We illustrate our method empirically by estimating a measurement error model using the CPS/SSR 1978 exact match file.

Residual-baseda posteriorierror analysis for the coupling of the Navier–Stokes and Darcy–Forchheimer equations

In this paper we consider a mixed variational formulation that have been recently proposed for the coupling of the Navier–Stokes and Darcy–Forchheimer equations, and derive, though in a non-standard sense, a reliable and efficient residual-baseda posteriorierror estimator suitable for an adaptive mesh-refinement method. For the reliability estimate, which holds with respect to the square root of the error estimator, we make use of the inf-sup condition and the strict monotonicity of the operators involved, a suitable Helmholtz decomposition in non-standard Banach spaces in the porous medium, local approximation properties of the Clément interpolant and Raviart–Thomas operator, and a smallness assumption on the data. In turn, inverse inequalities, the localization technique based on triangle-bubble and edge-bubble functions in localLpspaces, are the main tools for developing the efficiency analysis, which is valid for the error estimator itself up to a suitable additional error term. Finally, several numerical results confirming the properties of the estimator and illustrating the performance of the associated adaptive algorithm are reported.

Errors in the Dependent Variable of Quantile Regression Models

We study the consequences of measurement error in the dependent variable of random‐coefficients models, focusing on the particular case of quantile regression. The popular quantile regression estimator of Koenker and Bassett (1978) is biased if there is an additive error term. Approaching this problem as an errors‐in‐variables problem where the dependent variable suffers from classical measurement error, we present a sieve maximum likelihood approach that is robust to left‐hand‐side measurement error. After providing sufficient conditions for identification, we demonstrate that when the number of knots in the quantile grid is chosen to grow at an adequate speed, the sieve‐maximum‐likelihood estimator is consistent and asymptotically normal, permitting inference via bootstrapping. Monte Carlo evidence verifies our method outperforms quantile regression in mean bias and MSE. Finally, we illustrate our estimator with an application to the returns to education highlighting changes over time in the returns to education that have previously been masked by measurement‐error bias.

Cycle Lengths in Expanding Graphs

For a positive constant α a graph G on n vertices is called an α-expander if every vertex set U of size at most n/2 has an external neighborhood whose size is at least α|U|. We study cycle lengths in expanding graphs. We first prove that cycle lengths in α-expanders are well distributed. Specifically, we show that for every 0 < α ≤ 1 there exist positive constants n0, C and A = O(1/α) such that for every α-expander G on n≥n0 vertices and every integer $$\ell \in \left[ {C\log n,\tfrac{n}{C}} \right]$$ , G contains a cycle whose length is between $$\ell$$ and $$\ell$$ +A; the order of dependence of the additive error term A on α is optimal. Secondly, we show that every α-expander on n vertices contains $$\Omega \left( {\tfrac{{{\alpha ^3}}}{{\log (1/\alpha )}}} \right)$$ different cycle lengths. Finally, we introduce another expansion-type property, guaranteeing the existence of a linearly long interval in the set of cycle lengths. For β > 0 a graph G on n vertices is called a β-graph if every pair of disjoint sets of size at least βn are connected by an edge. We prove that for every $$\beta < 1/20$$ there exist positive constants $${b_1} = O\left( {\tfrac{1}{{\log (1 - \beta )}}} \right)$$ and b2 = O(β) such that every β-graph G on n vertices contains a cycle of length $$\ell$$ for every integer $$\ell \in \left[ {{b_1}\log n,\left( {1 - {b_2}} \right)n} \right]$$ ; the order of dependence of b1 and b2 on β is optimal.

A Review of Innovation-Based Methods to Jointly Estimate Model and Observation Error Covariance Matrices in Ensemble Data Assimilation

AbstractData assimilation combines forecasts from a numerical model with observations. Most of the current data assimilation algorithms consider the model and observation error terms as additive Gaussian noise, specified by their covariance matrices and , respectively. These error covariances, and specifically their respective amplitudes, determine the weights given to the background (i.e., the model forecasts) and to the observations in the solution of data assimilation algorithms (i.e., the analysis). Consequently, and matrices significantly impact the accuracy of the analysis. This review aims to present and to discuss, with a unified framework, different methods to jointly estimate the and matrices using ensemble-based data assimilation techniques. Most of the methods developed to date use the innovations, defined as differences between the observations and the projection of the forecasts onto the observation space. These methods are based on two main statistical criteria: 1) the method of moments, in which the theoretical and empirical moments of the innovations are assumed to be equal, and 2) methods that use the likelihood of the observations, themselves contained in the innovations. The reviewed methods assume that innovations are Gaussian random variables, although extension to other distributions is possible for likelihood-based methods. The methods also show some differences in terms of levels of complexity and applicability to high-dimensional systems. The conclusion of the review discusses the key challenges to further develop estimation methods for and . These challenges include taking into account time-varying error covariances, using limited observational coverage, estimating additional deterministic error terms, or accounting for correlated noise.

An Additive Approximation to Multiplicative Noise

Multiplicative noise models are often used instead of additive noise models in cases in which the noise variance depends on the state. Furthermore, when Poisson distributions with relatively small counts are approximated with normal distributions, multiplicative noise approximations are straightforward to implement. There are a number of limitations in the existing approaches to deal with multiplicative errors, such as positivity of the multiplicative noise term. The focus in this paper is on large dimensional (inverse) problems for which sampling-type approaches have too high computational complexity. In this paper, we propose an alternative approach utilising the Bayesian framework to carry out approximative marginalisation over the multiplicative error by embedding the statistics in an additive error term. The Bayesian framework allows the statistics of the resulting additive error term to be found based on the statistics of the other unknowns. As an example, we consider a deconvolution problem on random fields with different statistics of the multiplicative noise. Furthermore, the approach allows for correlated multiplicative noise. We show that the proposed approach provides feasible error estimates in the sense that the posterior models support the actual image.

Microlocally accurate solution of the inverse source problem of thermoacoustic tomography

We consider the inverse source problem of thermo- and photoacoustic tomography, with data registered on an open surface partially surrounding the source of acoustic waves. Under the assumption of constant speed of sound we develop an explicit non-iterative reconstruction procedure that recovers the Radon transform of the sought source, up to an infinitely smooth additive error term. The source then can be found by inverting the Radon transform. Our analysis is microlocal in nature and does not provide a norm estimate on the error in the so obtained image. However, numerical simulations show that this error is quite small in practical terms. We also present an asymptotically fast implementation of this procedure for the case when the data are given on a circular arc in 2D.

Boundedness of Massless Scalar Waves on Kerr Interior Backgrounds

We consider solutions of the massless scalar wave equation $\Box_g\psi=0$, without symmetry, on fixed subextremal Kerr backgrounds $(\mathcal M, g)$. It follows from previous analyses in the Kerr exterior that for solutions $\psi$ arising from sufficiently regular data on a two ended Cauchy hypersurface, the solution and its derivatives decay suitably fast along the event horizon $\mathcal H^+$. Using the derived decay rate, we show that $\psi$ is in fact uniformly bounded, $|\psi|\leq C$, in the black hole interior up to and including the bifurcate Cauchy horizon $\mathcal C\mathcal H^+$, to which $\psi$ in fact extends continuously. In analogy to our previous paper, [30], on boundedness of solutions to the massless scalar wave equation on fixed subextremal Reissner--Nordstr\"om backgrounds, the analysis depends on weighted energy estimates, commutation by angular momentum operators and application of Sobolev embedding. In contrast to the Reissner--Nordstr\"om case the commutation leads to additional error terms that have to be controlled.

Adaptive periodic event-triggered control for missile-target interception system with finite-horizon convergence

In this paper, the optimal control problem for finite-time missile-target interception systems is posed in a finite-horizon two-player zero-sum (ZS) differential game framework using a periodic event-triggered (PET) scheme. To solve the optimal control problem, a time-varying Hamilton-Jacobi-Issac (HJI) equation and a time-dependent cost function are constructed to deal with finite-horizon constraints, and an event-based periodic adaptive dynamic programming (ADP) algorithm is employed to find the Nash equilibrium solution for the designed HJI equation. Comparing with the traditional continuous event-triggered (ET) scheme, the proposed PET scheme only verifies the event-triggered conditions at periodic sampling instants, which reduces resource consumption in monitoring and excludes the Zeno behavior. A single critic neural network (CNN) is used to implement the proposed event-based optimal control algorithm, which reduces approximate errors bust also simplifies structures. Further, an additional error term is added in the designed weight updating law to such that the terminal constraint is also minimized over time. By resorting to Lyapunov function approach, some sufficient conditions are derived to achieve the uniformly ultimately bounded (UUB) of the ET closed-loop system and the estimation weight error of CNN. Finally, a missile-target interception system is introduced to illustrate the efficiency of the presented methods.

SRSC: Selective, Robust, and Supervised Constrained Feature Representation for Image Classification.

Feature representation learning, an emerging topic in recent years, has achieved great progress. Powerful learned features can lead to excellent classification accuracy. In this article, a selective and robust feature representation framework with a supervised constraint (SRSC) is presented. SRSC seeks a selective, robust, and discriminative subspace by transforming the original feature space into the category space. Particularly, we add a selective constraint to the transformation matrix (or classifier parameter) that can select discriminative dimensions of the input samples. Moreover, a supervised regularization is tailored to further enhance the discriminability of the subspace. To relax the hard zero-one label matrix in the category space, an additional error term is also incorporated into the framework, which can lead to a more robust transformation matrix. SRSC is formulated as a constrained least square learning (feature transforming) problem. For the SRSC problem, an inexact augmented Lagrange multiplier method (ALM) is utilized to solve it. Extensive experiments on several benchmark data sets adequately demonstrate the effectiveness and superiority of the proposed method. The proposed SRSC approach has achieved better performances than the compared counterpart methods.

Accelerated failure time models with log-concave errors

Summary We study accelerated failure time models in which the survivor function of the additive error term is log-concave. The log-concavity assumption covers large families of commonly used distributions and also represents the aging or wear-out phenomenon of the baseline duration. For right-censored failure time data, we construct semiparametric maximum likelihood estimates of the finite-dimensional parameter and establish the large sample properties. The shape restriction is incorporated via a nonparametric maximum likelihood estimator of the hazard function. Our approach guarantees the uniqueness of a global solution for the estimating equations and delivers semiparametric efficient estimates. Simulation studies and empirical applications demonstrate the usefulness of our method.

Bayesian methods for time series of count data

In this paper, we consider Bayesian methods for analyzing time series of count data under a Poisson regression model with a latent auto-regressive process embedded as an additive error term. We propose two different methods; the first method samples the latent variables one by one while the second method samples them jointly. The two methods are compared by simulation studies and an example employing real data. In terms of relative bias and root-mean-squared-errors, the two methods perform almost the same. However, the mixing performance of the first method is better than the second method for most of the simulation scenarios.