Source author record

John M. Noble

Explore this author’s publications, collaborators and research topics. This catalog record has not been claimed by its author.

ResearcherUnclaimed source record

math.PR Computation Machine Learning math.AP math.GM math.ST Statistics Theory

Catalog footprint

Research overview

6works

7topics

2close collaborators

Actions

Is this your work?

Open graph Browse papers

preprint2022arXiv

Lp Solutions for Stochastic Evolution Equation with Nonlinear Potential

This article considers the stochastic partial differential equation $\left\{ \begin{array}{l} u_t = \frac{1}{2} u_{xx} + u^γξu(0,.) = u_0 \end{array}\right.$ \noindent where $ξ$ is a space / time white noise Gaussian random field, $γ> 1$ and $u_0$ is a non-negative initial condition independent of $ξ$ satisfying $u_0 \geq 0, \qquad \lim_{n \rightarrow +\infty} \mathbb{E} \left [ \left (\int_{\mathbb{S}^1} u_0 (x)\wedge n dx \right)^2 \right ] = \mathbb{E} \left [ \left (\int_{\mathbb{S}^1} u_0 (x) dx \right)^2 \right ]< +\infty.$ \noindent The {\em space} variable is $x \in \mathbb{S}^1 = [0,1]$ with the identification $0 = 1$ . The definition of the stochastic term, taken in the sense of Walsh, will be made clear in the article. The result is that there exists a unique non-negative solution $u$ such that for all $α\in [0,1)$ , $\mathbb{E} \left [ \left( \int_0^\infty \int_{\mathbb{S}^1} u(t,x)^{2γ} dx dt \right)^{α/ 2 } \right ] \leq C( α) < + \infty.$ \noindent where the constant $C(α)$ arises in the Burkholder-Davis-Gundy inequality. The solution is also shown to satisfy $\mathbb{E} \left [ \int_0^T \left(\int_{\mathbb{S}^1} u (t,x)^p dx \right)^{α/ p} dt \right ] < +\infty \qquad \forall T < +\infty, \qquad p < +\infty, \qquad α\in \left (0, \frac{1}{2} \right ).$

preprint2015arXiv

A Prior Distribution over Directed Acyclic Graphs for Sparse Bayesian Networks

The main contribution of this article is a new prior distribution over directed acyclic graphs, which gives larger weight to sparse graphs. This distribution is intended for structured Bayesian networks, where the structure is given by an ordered block model. That is, the nodes of the graph are objects which fall into categories (or blocks); the blocks have a natural ordering. The presence of a relationship between two objects is denoted by an arrow, from the object of lower category to the object of higher category. The models considered here were introduced in Kemp et al. (2004) for relational data and extended to multivariate data in Mansinghka et al. (2006). The prior over graph structures presented here has an explicit formula. The number of nodes in each layer of the graph follow a Hoppe Ewens urn model. We consider the situation where the nodes of the graph represent random variables, whose joint probability distribution factorises along the DAG. We describe Monte Carlo schemes for finding the optimal aposteriori structure given a data matrix and compare the performance with Mansinghka et al. (2006) and also with the uniform prior.

preprint2014arXiv

Time Homogeneous Diffusion with drift and killing to meet a given marginal

This article gives conditions on a probability measure and drift field b such that for a given killing field k and a given time t > 0, there is function a such that there is a time homogeneous Markov process with infinitesimal generator a((1/2)d^2/dx^2 + b d/dx - k) which meets the given marginal at time t.

preprint2012arXiv

Time Homogeneous Diffusions with a Given Marginal at a Deterministic Time

In this article, it is proved that for any cumulative distribution function with compact support and a specified t > 0, there exists a diffusion martingale which has this law at time t. The article proves existence; no claims are made about uniqueness. After a discussion on strings and associated semigroups, the article gives a re-working of a standard approach to the problem of constructing an explicit discrete time martingale diffusion on a finite state space which, for a random geometrically distributed time that is independent of the diffusion, the law of the diffusion stopped at this random time has the prescribed law. This argument is developed, using a fixed point theorem, to determine conditions under which there is a discrete time martingale diffusion that has a prescribed law at an independent random time with negative binomial distribution. The step length for the time discretisation is then reduced and in the limit it is shown that for a finite state space, there exists a continuous time martingale diffusion such that $X_τ$ has law $μ$ , where $τ$ has a Gamma distribution. For fixed $t = E[τ]$ , the parameters of the Gamma distribution may be altered, reducing the coefficient of variation of $τ$ to zero, to show that there is a martingale diffusion $X$ such that $X_t$ has law $μ$ . The argument is then extended to obtain the result for any state space that is a bounded measurable subset of $R$ .

preprint2010arXiv

An Algorithm for Learning the Essential Graph

This article presents an algorithm for learning the essential graph of a Bayesian network. The basis of the algorithm is the Maximum Minimum Parents and Children algorithm developed by previous authors, with three substantial modifications. The MMPC algorithm is the first stage of the Maximum Minimum Hill Climbing algorithm for learning the directed acyclic graph of a Bayesian network, introduced by previous authors. The MMHC algorithm runs in two phases; firstly, the MMPC algorithm to locate the skeleton and secondly an edge orientation phase. The computationally expensive part is the edge orientation phase. The first modification introduced to the MMPC algorithm, which requires little additional computational cost, is to obtain the immoralities and hence the essential graph. This renders the edge orientation phase, the computationally expensive part, unnecessary, since the entire Markov structure that can be derived from data is present in the essential graph. Secondly, the MMPC algorithm can accept independence statements that are logically inconsistent with those rejected, since with tests for independence, a `do not reject' conclusion for a particular independence statement is taken as `accept' independence. An example is given to illustrate this and a modification is suggested to ensure that the conditional independence statements are logically consistent. Thirdly, the MMHC algorithm makes an assumption of faithfulness. An example of a data set is given that does not satisfy this assumption and a modification is suggested to deal with some situations where the assumption is not satisfied. The example in question also illustrates problems with the `faithfulness' assumption that cannot be tackled by this modification.

preprint2010arXiv

On the Stochastic Burgers Equation and the Axiom of Choice

The equivalence of the Choice Axiom and Tychonov compactness was proved by Kelley in 1950. Tychonov compactness is required to prove existence of the minimiser of an action functional under standard hypotheses. Solutions to the inviscid Burgers equation may be constructed by considering the minimising trajectories of associated action functionals. This construction has been used in the literature to construct an invariant measure for the stochastic inviscid Burgers equation. This article compares results that may be obtained with and without using this representation and draws conclusions about the Choice Axiom.

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.