Multidimensional Gaussian Process Research Articles

Symplectic Runge–Kutta methods for Hamiltonian systems driven by Gaussian rough paths

We consider Hamiltonian systems driven by multi-dimensional Gaussian processes in rough path sense, which include fractional Brownian motions with Hurst parameter H∈(1/4,1/2]. We prove that the phase flow preserves the symplectic structure almost surely and this property could be inherited by symplectic Runge–Kutta methods, which are implicit methods in general. If the vector fields satisfy some smoothness and boundedness conditions, we obtain the pathwise convergence rates of Runge–Kutta methods. When vector fields are linear, we get the solvability of the midpoint scheme for skew symmetric cases, and obtain its pathwise convergence rate. Numerical experiments verify our theoretical analysis.

Discrete rough paths and limit theorems

In this article, we consider limit theorems for some weighted type random sums (or discrete rough integrals). We introduce a general transfer principle from limit theorems for unweighted sums to limit theorems for weighted sums via rough path techniques. As a by-product, we provide a natural explanation of the various new asymptotic behaviors in contrast with the classical unweighted random sum case. We apply our principle to derive some weighted type Breuer-Major theorems, which generalize previous results to random sums that do not have to be in a finite sum of chaos. In this context, a Breuer-Major type criterion in notion of Hermite rank is obtained. We also consider some applications to realized power variations and to Ito's formulas in law. In the end, we study the asymptotic behavior of weighted quadratic variations for some multi-dimensional Gaussian processes.

Generalized p value for multivariate Gaussian stochastic processes in continuous time

We construct a Generalized p value for testing statistical hypotheses on the comparison of mean vectors in the sequential observation of two continuous time multidimensional Gaussian processes. The mean vectors depend linearly on two multidimensional parameters and with different conditions about their covariance structures. The invariance of the generalized p value considered is proved under certain linear transformations. We report results of a simulation study showing power and errors probabilities for them. Finally, we apply our results to a real data set.

Regularized nonparametric Volterra kernel estimation

In this paper, the regularization approach introduced recently for nonparametric estimation of linear systems is extended to the estimation of nonlinear systems modeled as Volterra series. The kernels of order higher than one, representing higher dimensional impulse responses in the series, are considered to be realizations of multidimensional Gaussian processes. Based on this, prior information about the structure of the Volterra kernel is introduced via an appropriate penalization term in the least squares cost function. It is shown that the proposed method is able to deliver accurate estimates of the Volterra kernels even in the case of a small amount of data points.

Sensitivitiesviarough paths

Motivated by a problematic coming from mathematical finance, this paper is devoted to existing and additional results of continuity and differentiability of the It\^o map associated to rough differential equations. These regularity results together with Malliavin calculus are applied to sensitivities analysis for stochastic differential equations driven by multidimensional Gaussian processes with continuous paths, especially fractional Brownian motions. Precisely, in that framework, results on computation of greeks for It\^o's stochastic differential equations are extended. An application in mathematical finance, and simulations, are provided.

On Stratonovich and Skorohod stochastic calculus for Gaussian processes

In this article, we derive a Stratonovich and Skorohod-type change of variables formula for a multidimensional Gaussian process with low Hölder regularity $\gamma$ (typically $\gamma \le1/4$). To this aim, we combine tools from rough paths theory and stochastic analysis.

Multi-output separable Gaussian process: Towards an efficient, fully Bayesian paradigm for uncertainty quantification

Computer codes simulating physical systems usually have responses that consist of a set of distinct outputs (e.g., velocity and pressure) that evolve also in space and time and depend on many unknown input parameters (e.g., physical constants, initial/boundary conditions, etc.). Furthermore, essential engineering procedures such as uncertainty quantification, inverse problems or design are notoriously difficult to carry out mostly due to the limited simulations available. The aim of this work is to introduce a fully Bayesian approach for treating these problems which accounts for the uncertainty induced by the finite number of observations. Our model is built on a multi-dimensional Gaussian process that explicitly treats correlations between distinct output variables as well as space and/or time. The proper use of a separable covariance function enables us to describe the huge covariance matrix as a Kronecker product of smaller matrices leading to efficient algorithms for carrying out inference and predictions. The novelty of this work, is the recognition that the Gaussian process model defines a posterior probability measure on the function space of possible surrogates for the computer code and the derivation of an algorithmic procedure that allows us to sample it efficiently. We demonstrate how the scheme can be used in uncertainty quantification tasks in order to obtain error bars for the statistics of interest that account for the finite number of observations.

The Gaussian approximation for generalized Friedman’s urn model with heterogeneous and unbalanced updating

The Friedman’s urn model is a popular urn model which is widely used in many disciplines. In particular, it is extensively used in treatment allocation schemes in clinical trials. Its asymptotic properties have been studied by many researchers. In literature, it is usually assumed that the expected number of balls added at each stage is a constant in despite of what type of balls are selected, that is, the updating of the urn is assumed to be balanced. When it is not, the asymptotic property of the Friedman’s urn model is stated in the book of Hu and Rosenberger (2006) as one of open problems in the area of adaptive designs. In this paper, we show that both the urn composition process and the allocation proportion process can be approximated by a multi-dimensional Gaussian process almost surely for a general multi-color Friedman type urn model with heterogeneous and unbalanced updating. The Gaussian process is a solution of a stochastic differential equation. As an application, we obtain the asymptotic properties including the asymptotic normality and the exact law of the iterated logarithm.

Smoothness of the density for solutions to Gaussian rough differential equations

We consider stochastic differential equations of the form $dY_t=V(Y_t)\,dX_t+V_0(Y_t)\,dt$ driven by a multi-dimensional Gaussian process. Under the assumption that the vector fields $V_0$ and $V=(V_1,\ldots,V_d)$ satisfy H\"{o}rmander's bracket condition, we demonstrate that $Y_t$ admits a smooth density for any $t\in(0,T]$, provided the driving noise satisfies certain nondegeneracy assumptions. Our analysis relies on relies on an interplay of rough path theory, Malliavin calculus and the theory of Gaussian processes. Our result applies to a broad range of examples including fractional Brownian motion with Hurst parameter $H>1/4$, the Ornstein-Uhlenbeck process and the Brownian bridge returning after time $T$.

Sequential Analysis of Censored Data with Linear Transformation Models

Sequential tests have been used commonly in clinical trials to compare treatments. For sequential analysis of right-censored survival data with covariate adjustment, several different methods have been studied based on either Cox proportional hazards model or accelerated failure time model. Here we propose a test process based on linear transformation models for staggered entry data. The proposed test process is motivated by Chen et al.'s (2002) estimating equations for linear transformation models. We show that the test process can be approximated by a mean 0 multidimensional Gaussian process. A consistent estimator of its covariance matrix function is provided. For given interim analysis time points, a repeated significant test is developed based on the boundaries procedure proposed by Slud and Wei (1982). Numerical studies show that the proposed test process performs well.

Processes with Block-Associated Increments

This paper is motivated by relations between association and independence of random variables. It is well known that, for real random variables, independence implies association in the sense of Esary, Proschan and Walkup (1967), while, for random vectors, this simple relationship breaks. We modify the notion of association in such a way that any vector-valued process with independent increments also has associated increments in the new sense - association between blocks. The new notion is quite natural and admits nice characterization for some classes of processes. In particular, using the covariance interpolation formula due to Houdré, Pérez-Abreu and Surgailis (1998), we show that within the class of multidimensional Gaussian processes, block association of increments is equivalent to supermodularity (in time) of the covariance functions. We also define corresponding versions of weak association, positive association, and negative association. It turns out that the central limit theorem for weakly associated random vectors due to Burton, Dabrowski and Dehling (1986) remains valid, if the weak association is relaxed to the weak association between blocks.

Processes with Block-Associated Increments

This paper is motivated by relations between association and independence of random variables. It is well known that, for real random variables, independence implies association in the sense of Esary, Proschan and Walkup (1967), while, for random vectors, this simple relationship breaks. We modify the notion of association in such a way that any vector-valued process with independent increments also has associated increments in the new sense - association between blocks. The new notion is quite natural and admits nice characterization for some classes of processes. In particular, using the covariance interpolation formula due to Houdré, Pérez-Abreu and Surgailis (1998), we show that within the class of multidimensional Gaussian processes, block association of increments is equivalent to supermodularity (in time) of the covariance functions. We also define corresponding versions of weak association, positive association, and negative association. It turns out that the central limit theorem for weakly associated random vectors due to Burton, Dabrowski and Dehling (1986) remains valid, if the weak association is relaxed to the weak association between blocks.

Extremes of multidimensional Gaussian processes

This paper considers extreme values attained by a centered, multidimensional Gaussian process X ( t ) = ( X 1 ( t ) , … , X n ( t ) ) minus drift d ( t ) = ( d 1 ( t ) , … , d n ( t ) ) , on an arbitrary set T . Under mild regularity conditions, we establish the asymptotics of log P ( ∃ t ∈ T : ⋂ i = 1 n { X i ( t ) − d i ( t ) > q i u } ) , for positive thresholds q i > 0 , i = 1 , … , n and u → ∞ . Our findings generalize and extend previously known results for the single-dimensional and two-dimensional cases. A number of examples illustrate the theory.

Densities for rough differential equations under Hörmander’s condition

We consider stochastic differential equations dY = V (Y) dX driven by a multidimensional Gaussian process X in the rough path sense [T. Lyons, Rev. Mat. Iberoamericana 14, (1998), 215-310]. Using Malliavin Calculus we show that Y t admits a density for t ∈ (0, T] provided (i) the vector fields V = (V 1 , ... , V d ) satisfy Hormander's condition and (ii) the Gaussian driving signal X satisfies certain conditions. Examples of driving signals include fractional Brownian motion with Hurst parameter H > 1/4, the Brownian bridge returning to zero after time T and the Ornstein-Uhlenbeck process.

Asymptotic behaviour of a test statistic for detection of change in mean of vectors

The paper gives an asymptotic distribution of a test statistic for detecting a change in a mean of random vectors with dependent components. The studied test statistic has a form of a maximum of a square Euclidean norms of vectors with components being standardized partial cumulative sums of deviations from means. The limit distribution was obtained using a result of Piterbarg [1994. High deviations for multidimensional stationary Gaussian processes with independent components. In: Zolotarev, V.M. (Ed.), Stability Problems for Stochastic Models, pp. 197–210].

The Gaussian approximation for multi-color generalized Friedman’s urn model

The generalized Friedman’s urn model is a popular urn model which is widely used in many disciplines. In particular, it is extensively used in treatment allocation schemes in clinical trials. In this paper, we show that both the urn composition process and the allocation proportion process can be approximated by a multi-dimensional Gaussian process almost surely for a multi-color generalized Friedman’s urn model with both homogeneous and non-homogeneous generating matrices. The Gaussian process is a solution of a stochastic differential equation. This Gaussian approximation is important for the understanding of the behavior of the urn process and is also useful for statistical inferences. As an application, we obtain the asymptotic properties including the asymptotic normality and the law of the iterated logarithm for a multi-color generalized Friedman’s urn model as well as the randomized-play-the-winner rule as a special case.

Gaussian point processes and two-by-two random matrix theory

The statistics of the multidimensional Gaussian point process are discussed in connection with the spacing statistics of eigenvalues of 2x2 random matrices. We consider the three-dimensional Gaussian point process when two of the coordinates of a point are randomly chosen from a Gaussian distribution having a mean of zero and a variance of sigma;{2}=1 but the third coordinate is chosen from a Gaussian distribution having a variance in the range of 0< or =sigma_{3};{2}< or =1 . The probability density function associated with a random point being at a distance r from the origin is shown to be closely related to the nearest-neighbor spacing distribution of eigenvalues coming from an ensemble of 2x2 matrices defined by the French-Kota-Pandey-Mehta two-matrix model of random matrix theory. An elementary explanation of this result is given.

Collective Fluctuations in Ordered Fluids Investigated by Two-Dimensional Electron−Electron Double Resonance Spectroscopy

Two-dimensional electron-electron double resonance (2D-ELDOR) is a technique that is sensitive to the dynamical processes affecting spin labels in complex fluid environments. In ordered fluids, such as membrane vesicles, the 2D-ELDOR experiment is affected by the molecular tumbling in the locally ordered environment. This motion occurs on two different time scales, the faster molecular motion relative to the local director, and the slower collective fluctuations of the director field. In the experimental study of Patyal, Crepeau, and Freed (Biophys. J. 1997, 73, 2201), it was found that the widths of the autopeaks of the 2D-ELDOR spectrum increased as a function of the mixing time. In the present work, a theory is developed for the effects of director fluctuations on the autopeaks in the 2D-ELDOR experiment by employing an analytical solution of the stochastic Liouville equation for which the director field is treated as a multidimensional Gaussian process, as previously developed by Frezzato, Kothe, and Moro (J. Phys. Chem. B 2001, 105, 1281 and J. Phys. Chem. B 2004, 108, 9505). Good agreement is found between theory and experiment, notably the only adjustable parameter is k, the bending elastic modulus of the membrane. The values of k = 11 x 10(-20) J for 1,2-dipalmitoyl-sn-glycero-phosphatidylcholine (DPPC) vesicles and k = 15 x 10(-20) J for DPPC/gramicidin A (5:1) vesicles, both at 45 degrees C, were found from the analysis and agree well with previous related measurements by other physical techniques. This establishes 2D-ELDOR as a useful technique to study the elastic properties of biological membranes.

On Behrens–Fisher problem for continuous time Gaussian processes

Our setting is the sequential observation of two continuous time multidimensional Gaussian processes whose mean vectors depend linearly on two multidimensional parameters and with different conditions about their covariance structures that will always include nuisance parameters. We analyze the Behrens–Fisher problem of comparing both parameters by means of a confidence set for their difference, with given confidence level and diameter. The random time needed to achieve this goal is also inspected.

Inference about a secondary process following a sequential trial

SUMMARY We consider the following sequential testing problem. A group-sequential or fullysequential test is carried out for a primary parameter, using a score process or an effective score process to eliminate nuisance parameters. After stopping, the possibility of additional parameters is considered, and appropriate tests and estimators are desired that recognise the sequential stopping rule. We formulate an asymptotic multi-dimensional Gaussian process form of such problems, and then construct tests and confidence procedures. Optimality conditions are given, and an example is summarised.