Gaussian channels and the optimal coding

For the Gaussian channel Y(t) = Φ(ξ(s), Y(s); s ≦ t) + X(t), the mutual information I(ξ, Y) between the message ξ(·) and the output Y(·) is evaluated, where X(·) is a Gaussian noise. Furthermore, the optimal coding under average power constraints is constructed.     

On the capacity of the continuous time Gaussian channel with feedback

We discuss the capacity of the Gaussian channel with feedback. In general it is not easy to give an explicit formula for the capacity of a Gaussian channel, unless the channel is without feedback or a white Gaussian channel. We consider the case where a constraint, given in terms of the covariance functions of the input processes, is imposed on the input processes. It is shown that the capacity of the Gaussian channel can be achieved by transmitting a Gaussian message and using additive linear feedback.     

Coding theory in white Gaussian channel with feedback

The message m = {m(t)} is a Gaussian process that is to be transmitted through the white Gaussian channel with feedback: $\text{Y(t) = ∫}{}_0{}^{\mathrm{t}}\text{F(s, Y}{}_0{}^{\mathrm{s}}\text{, m)ds + W(t)}$. Under the average power constraint, $\text{E[F}{}^2\text{(s, Y}{}_0{}^{\mathrm{s}}\text{, m)] ≤ P}{}_0$, we construct causally the optimal coding, in the sense that the mutual information I_t(m, Y) between the message m and the channel output Y (up to t) is maximized. The optimal coding is presented by $\text{Y(t) = ∫}{}_0{}^{\mathrm{t}}\text{A(s)[m(s) ?}\text{m}\text{?}\text{(s)] ds + W(t)}$, where $\text{m}\text{?}\text{(s) = E[m(s) ¦ Y(u), 0 ≤ u ≤ s]}$ and A(s) is a positive function such that $\text{A}{}^2\text{(s) E |m(s) ?}\text{m}\text{?}\text{(s)|}{}^2\text{= P}{}_0$.     

Martingale convergence,series expansion of Gaussian elements,and strong law of large numbers in Fréchet spaces

The main result of this paper is the derivation of a convergence theorem for certain martingales with values in a separable Fréchet space F. It is shown that this result includes a well known theorem due to Chatterji. Moreover, the series expansion of zero-mean Gaussian elements with values in F and the strong law of large numbers for i.i.d. F-valued random elements also follow as applications of the main theorem.     

Optimal transmission of Gaussian signals through a feedback channel

Using the methodology and results of the theory of filtering of conditionally Gaussian processes, the optimal schemes of transmission of Gaussian signals through the noisy feedback channel are constructed under the new power conditions.     

Sharp asymptotics of the Kolmogorov entropy for Gaussian measures

Let U denote the unit ball of the Cameron-Martin space of a Gaussian measure on a Hilbert space. The sharp asymptotics for the Kolmogorov (metric) entropy numbers of U is derived. The condition imposed is regular variation of the eigenvalues of the covariance operator. A consequence is a precise link including constants to the functional quantization problem.     

On Gaussian Processes Equivalent in Law to Fractional Brownian Motion

We consider Gaussian processes that are equivalent in law to the fractional Brownian motion and their canonical representations. We prove a Hitsuda type representation theorem for the fractional Brownian motion with Hurst index H1/2. For the case H>1/2 we show that such a representation cannot hold. We also consider briefly the connection between Hitsuda and Girsanov representations. Using the Hitsuda representation we consider a certain special kind of Gaussian stochastic equation with fractional Brownian motion as noise.     

Large deviations of infinite intersections of events in Gaussian processes

Consider events of the form {_Zs≥ζ(s),s∈S}$\{Z_s\ge \zeta \left(s\right),s\in S\}$, where Z$Z$ is a continuous Gaussian process with stationary increments, ζ$\zeta$ is a function that belongs to the reproducing kernel Hilbert space R$R$ of process Z$Z$, and S⊂R$S\subset \mathbb{R}$ is compact. The main problem considered in this paper is identifying the function β^∗∈R${\beta }^{\ast }\in R$ satisfying β^∗(s)≥ζ(s)${\beta }^{\ast }\left(s\right)\ge \zeta \left(s\right)$ on S$S$ and having minimal R$R$-norm. The smoothness (mean square differentiability) of Z$Z$ turns out to have a crucial impact on the structure of the solution. As examples, we obtain the explicit solutions when ζ(s)=s$\zeta \left(s\right)=s$ for s∈[0,1]$s\in [0,1]$ and Z$Z$ is either a fractional Brownian motion or an integrated Ornstein–Uhlenbeck process.     

Existence of the linear prediction for Banach space valued Gaussian processes

The correspondence between Gaussian stochastic processes with values in a Banach space E and cylindrical processes which are related to them is studied. It is shown that the linear prediction of an E-valued Gaussian process is an E-valued random variable as well as the spectral measure of an E-valued Gaussian stationary process is a Gaussian random measure.     

Interaction information and Markov chain

The problem of multivariate information analysis is considered. First, the interaction information in each dimension is defined analogously according to McGill [4] and then applied to Markov chains. The property of interaction information zero deeply relates to a certain class of weakly dependent random variables. For homogeneous, recurrent Markov chains with m states, m ≥ n ≥3, the zero criterion of n-dimensional interaction information is achieved only by (n ? 2)-dependent Markov chains, which are generated by some nilpotent matrices. Further for Gaussian Markov chains, it gives the decomposition rule of the variables into mutually correlated subchains.     

Gaussian processes with biconvex covariances

Let R(s, t) be a continuous, nonnegative, real valued function on a ≤ s ≤ t ≤ b. Suppose $\text{?R}\text{?s}\text{≥ 0}$, $\text{?R}\text{?t}\text{≤ 0}$, and $\text{?}{}^2\text{R}\text{?t}\text{?t ≤ 0}$ in the interior of the domain. Then the extension of R to a symmetric function on [a, b] × [a, b] is a covariance function. Such a covariance is called biconvex. Let X(t) be a Gaussian process with mean 0 and biconvex covariance. X has a representation as a sum of simple moving averages of white noises on the line and plane. The germ field of X at every point t is generated by X(t) alone. X is locally nondeterministic. Under an additional assumption involving the partial derivatives of R near the diagonal, the local time of the sample function exists and is jointly continuous almost surely, so that the sample function is nowhere differentiable.     

On the Equivalence of Multiparameter Gaussian Processes

Our purpose is to characterize the multiparameter Gaussian processes, that is Gaussian sheets, that are equivalent in law to the Brownian sheet and to the fractional Brownian sheet. We survey multiparameter analogues of the Hitsuda, Girsanov and Shepp representations. As an application, we study a special type of stochastic equation with linear noise.      

Prediction for two processes and the nehari problem

We exploit an analogy between the trigonometric moment problem and prediction theory for a stationary stochastic process. Extending this theory, we show how to use correlations between two processes to predict one from the other. In turn, this gives rise to a simple and unified treatment of the Caratheodory and Nehari moment problems.     

Equivalence of measures for some class of Gaussian random fields

We consider two Gaussian measures P₁ and P₂ on (C(G), $\text{B}$) with zero expectations and covariance functions R₁(x, y) and R₂(x, y) respectively, where R_ν(x, y) is the Green's function of the Dirichlet problem for some uniformly strongly elliptic differential operator A^(ν) of order $\text{2m, m ≥ [}\text{d}\text{2}\text{] + 1}$, on a bounded domain G in $\text{R}$^d (ν = 1, 2). It is shown that if the order of A⁽²⁾ ? A⁽¹⁾ is at most $\text{2m ? [}\text{d}\text{2}\text{] ? 1}$, then P₁ and P₂ are equivalent, while if the order is greater than $\text{2m ? [}\text{d}\text{2}\text{] ? 1}$, then P₁ and P₂ are not always equivalent.     

算子方程基本问题解的再生核稀疏表示

在一个Hilbert空间中通过内积核定义的线性算子对应一个自然的再生核Hilbert空间结构.本文将称其为H-HK结构.这个结构本身内蕴一个基方法,可以解答线性算子的若干最基本的问题,包括确定或刻画其值域空间、解算子方程及解Moore-Penrose伪-(广义-)逆算子问题.在对已存在结果的简要综述之后,本文的目的是建立H-HK结构下的预正交自适应Fourier分解(pre-orthogonal adaptive Fourier decomposition,POAFD)算法.在这个方法之下导出上述3个问题的解的稀疏表示.在逐次跟踪匹配的优化方法论中POAFD的优选原理保证了它在理论上和实用上的最优性.它也具有算法上的可行性.所提供的方法可有效地应用于具体实际问题,包括信号与图像重构、常微分方程、偏微分方程和优化问题的数值解等.     

Equivalence of two notions of continuity for stationary continuous-time information sources

The notion of continuity for continuous-time information sources which was introduced by Pinsker has found numerous applications in information theory. Continuity in probability is an important concept in the theory of continuous-time stochastic processes. It is shown that these two forms of continuity are equivalent for stationary processes whose state space is a separable metric space.     

The Hausdorff dimension of multivariate operator-self-similar Gaussian random fields

Let $\{X\left(t\right):t\in {\mathbb{R}}^d\}$ be a multivariate operator-self-similar random field with values in ${\mathbb{R}}^m$. Such fields were introduced in [22] and satisfy the scaling property $\{X\left(c^{E}t\right):t\in {\mathbb{R}}^d\}\stackrel{\mathrm{d}}{=}\{c^{D}X\left(t\right):t\in {\mathbb{R}}^d\}$ for all $c>0$, where $E$ is a $d\times d$ real matrix and $D$ is an $m\times m$ real matrix. We solve an open problem in [22] by calculating the Hausdorff dimension of the range and graph of a trajectory over the unit cube $K={[0,1]}^d$ in the Gaussian case. In particular, we enlighten the property that the Hausdorff dimension is determined by the real parts of the eigenvalues of $E$ and $D$ as well as the multiplicity of the eigenvalues of $E$ and $D$.     

A test for the independence of two Gaussian processes

A bivariate Gaussian process with mean 0 and covariance

$\text{Σ(s, t, p)=}\text{Σ}{}_{11}\text{(s, t)}\text{ρΣ}{}_{12}\text{(s, t)}\text{ρΣ}{}_{21}\text{(s, t)}\text{Σ}{}_{22}\text{(s, t)}$

is observed in some region Ω of R′, where {Σ_ij(s,t)} are given functions and p an unknown parameter. A test of H₀: p = 0, locally equivalent to the likelihood ratio test, is given for the case when Ω consists of p points. An unbiased estimate of p is given. The case where Ω has positive (but finite) Lebesgue measure is treated by spreading the p points evenly over Ω and letting p → ∞. Two distinct cases arise, depending on whether Δ_2,p, the sum of squares of the canonical correlations associated with Σ(s, t, 1) on $\text{Ω}{}^2$, remains bounded. In the case of primary interest as p → ∞, Δ_2,p → ∞, in which case $\text{p}\text{?}$ converges to p and the power of the one-sided and two-sided tests of H₀ tends to 1. (For example, this case occurs when Σ_ij(s, t) ≡ Σ₁₁(s, t).)     

Average sampling and reconstruction in a reproducing kernel subspace of homogeneous type space

In this paper, we first introduce a reproducing kernel subspace of , where is a homogeneous type space. Then we consider average sampling and reconstruction of signals in the reproducing kernel subspace of . We show that signals in the reproducing kernel subspace of could be stably reconstructed from its average samples taken on a relatively‐separated set with small gap. Exponential convergence is established for the iterative approximation‐projection reconstruction algorithm.