arXiv subjects

Ben Calderhead

Publications and source records attributed to Ben Calderhead.

8 recordsLinked to original sources

arXiv2019

Efficient structure learning with automatic sparsity selection for causal graph processes

We propose a novel algorithm for efficiently computing a sparse directed adjacency matrix from a group of time series following a causal graph process. Our solution is scalable for both dense and sparse graphs and automatically selects the LASSO coefficient to obtain an appropriate number of edges in the adjacency matrix. Current state-of-the-art approaches rely on sparse-matrix-computation libraries to scale, and either avoid automatic selection of the LASSO penalty coefficient or rely on the prediction mean squared error, which is not directly related to the correct number of edges. Instead, we propose a cyclical coordinate descent algorithm that employs two new non-parametric error metrics to automatically select the LASSO coefficient. We demonstrate state-of-the-art performance of our algorithm on simulated stochastic block models and a real dataset of stocks from the S\&P$500$.

stat.ML↗

arXiv2019

Efficient computation of mean reverting portfolios using cyclical coordinate descent

The econometric challenge of finding sparse mean reverting portfolios based on a subset of a large number of assets is well known. Many current state-of-the-art approaches fall into the field of co-integration theory, where the problem is phrased in terms of an eigenvector problem with sparsity constraint. Although a number of approximate solutions have been proposed to solve this NP-hard problem, all are based on relatively simple models and are limited in their scalability. In this paper we leverage information obtained from a heterogeneous simultaneous graphical dynamic linear model (H-SGDLM) and propose a novel formulation of the mean reversion problem, which is phrased in terms of a quasi-convex minimisation with a normalisation constraint. This new formulation allows us to employ a cyclical coordinate descent algorithm for efficiently computing an exact sparse solution, even in a large universe of assets, while the use of H-SGDLM data allows us to easily control the required level of sparsity. We demonstrate the flexibility, speed and scalability of the proposed approach on S\&P$500$, FX and ETF futures data.

q-fin.PM↗

arXiv2019

A Dynamic Bayesian Model for Interpretable Decompositions of Market Behaviour

We propose a heterogeneous simultaneous graphical dynamic linear model (H-SGDLM), which extends the standard SGDLM framework to incorporate a heterogeneous autoregressive realised volatility (HAR-RV) model. This novel approach creates a GPU-scalable multivariate volatility estimator, which decomposes multiple time series into economically-meaningful variables to explain the endogenous and exogenous factors driving the underlying variability. This unique decomposition goes beyond the classic one step ahead prediction; indeed, we investigate inferences up to one month into the future using stocks, FX futures and ETF futures, demonstrating its superior performance according to accuracy of large moves, longer-term prediction and consistency over time.

q-fin.CP↗

arXiv2018

Quasi Markov Chain Monte Carlo Methods

Quasi-Monte Carlo (QMC) methods for estimating integrals are attractive since the resulting estimators typically converge at a faster rate than pseudo-random Monte Carlo. However, they can be difficult to set up on arbitrary posterior densities within the Bayesian framework, in particular for inverse problems. We introduce a general parallel Markov chain Monte Carlo (MCMC) framework, for which we prove a law of large numbers and a central limit theorem. In that context, non-reversible transitions are investigated. We then extend this approach to the use of adaptive kernels and state conditions, under which ergodicity holds. As a further extension, an importance sampling estimator is derived, for which asymptotic unbiasedness is proven. We consider the use of completely uniformly distributed (CUD) numbers within the above mentioned algorithms, which leads to a general parallel quasi-MCMC (QMCMC) methodology. We prove consistency of the resulting estimators and demonstrate numerically that this approach scales close to $n^{-2}$ as we increase parallelisation, instead of the usual $n^{-1}$ that is typical of standard MCMC algorithms. In practical statistical models we observe multiple orders of magnitude improvement compared with pseudo-random methods.

math.ST↗

arXiv2018

Implicit Probabilistic Integrators for ODEs

We introduce a family of implicit probabilistic integrators for initial value problems (IVPs), taking as a starting point the multistep Adams-Moulton method. The implicit construction allows for dynamic feedback from the forthcoming time-step, in contrast to previous probabilistic integrators, all of which are based on explicit methods. We begin with a concise survey of the rapidly-expanding field of probabilistic ODE solvers. We then introduce our method, which builds on and adapts the work of Conrad et al. (2016) and Teymur et al. (2016), and provide a rigorous proof of its well-definedness and convergence. We discuss the problem of the calibration of such integrators and suggest one approach. We give an illustrative example highlighting the effect of the use of probabilistic integrators - including our new method - in the setting of parameter inference within an inverse problem.

stat.ME↗

arXiv2016

Probabilistic Linear Multistep Methods

We present a derivation and theoretical investigation of the Adams-Bashforth and Adams-Moulton family of linear multistep methods for solving ordinary differential equations, starting from a Gaussian process (GP) framework. In the limit, this formulation coincides with the classical deterministic methods, which have been used as higher-order initial value problem solvers for over a century. Furthermore, the natural probabilistic framework provided by the GP formulation allows us to derive probabilistic versions of these methods, in the spirit of a number of other probabilistic ODE solvers presented in the recent literature. In contrast to higher-order Runge-Kutta methods, which require multiple intermediate function evaluations per step, Adams family methods make use of previous function evaluations, so that increased accuracy arising from a higher-order multistep approach comes at very little additional computational cost. We show that through a careful choice of covariance function for the GP, the posterior mean and standard deviation over the numerical solution can be made to exactly coincide with the value given by the deterministic method and its local truncation error respectively. We provide a rigorous proof of the convergence of these new methods, as well as an empirical investigation (up to fifth order) demonstrating their convergence rates in practice.

math.NA↗

arXiv2013

Bayesian Solution Uncertainty Quantification for Differential Equations

We explore probability modelling of discretization uncertainty for system states defined implicitly by ordinary or partial differential equations. Accounting for this uncertainty can avoid posterior under-coverage when likelihoods are constructed from a coarsely discretized approximation to system equations. A formalism is proposed for inferring a fixed but a priori unknown model trajectory through Bayesian updating of a prior process conditional on model information. A one-step-ahead sampling scheme for interrogating the model is described, its consistency and first order convergence properties are proved, and its computational complexity is shown to be proportional to that of numerical explicit one-step solvers. Examples illustrate the flexibility of this framework to deal with a wide variety of complex and large-scale systems. Within the calibration problem, discretization uncertainty defines a layer in the Bayesian hierarchy, and a Markov chain Monte Carlo algorithm that targets this posterior distribution is presented. This formalism is used for inference on the JAK-STAT delay differential equation model of protein dynamics from indirectly observed measurements. The discussion outlines implications for the new field of probabilistic numerics.

stat.ME↗

arXiv2009

Riemannian Manifold Hamiltonian Monte Carlo

The paper proposes a Riemannian Manifold Hamiltonian Monte Carlo sampler to resolve the shortcomings of existing Monte Carlo algorithms when sampling from target densities that may be high dimensional and exhibit strong correlations. The method provides a fully automated adaptation mechanism that circumvents the costly pilot runs required to tune proposal densities for Metropolis-Hastings or indeed Hybrid Monte Carlo and Metropolis Adjusted Langevin Algorithms. This allows for highly efficient sampling even in very high dimensions where different scalings may be required for the transient and stationary phases of the Markov chain. The proposed method exploits the Riemannian structure of the parameter space of statistical models and thus automatically adapts to the local manifold structure at each step based on the metric tensor. A semi-explicit second order symplectic integrator for non-separable Hamiltonians is derived for simulating paths across this manifold which provides highly efficient convergence and exploration of the target density. The performance of the Riemannian Manifold Hamiltonian Monte Carlo method is assessed by performing posterior inference on logistic regression models, log-Gaussian Cox point processes, stochastic volatility models, and Bayesian estimation of parameter posteriors of dynamical systems described by nonlinear differential equations. Substantial improvements in the time normalised Effective Sample Size are reported when compared to alternative sampling approaches. Matlab code at \url{http://www.dcs.gla.ac.uk/inference/rmhmc} allows replication of all results.

stat.CO↗