Showing 1–7 of 7 results for author: O'Kane, T

Bayesian Vector AutoRegression with Factorised Granger-Causal Graphs

Authors: He Zhao, Vassili Kitsios, Terence J. O'Kane, Edwin V. Bonilla

Abstract: We study the problem of automatically discovering Granger causal relations from observational multivariate time-series data.Vector autoregressive (VAR) models have been time-tested for this problem, including Bayesian variants and more recent developments using deep neural networks. Most existing VAR methods for Granger causality use sparsity-inducing penalties/priors or post-hoc thresholds to int… ▽ More We study the problem of automatically discovering Granger causal relations from observational multivariate time-series data.Vector autoregressive (VAR) models have been time-tested for this problem, including Bayesian variants and more recent developments using deep neural networks. Most existing VAR methods for Granger causality use sparsity-inducing penalties/priors or post-hoc thresholds to interpret their coefficients as Granger causal graphs. Instead, we propose a new Bayesian VAR model with a hierarchical factorised prior distribution over binary Granger causal graphs, separately from the VAR coefficients. We develop an efficient algorithm to infer the posterior over binary Granger causal graphs. Comprehensive experiments on synthetic, semi-synthetic, and climate data show that our method is more uncertainty aware, has less hyperparameters, and achieves better performance than competing approaches, especially in low-data regimes where there are less observations. △ Less

Submitted 23 May, 2024; v1 submitted 5 February, 2024; originally announced February 2024.

Variational DAG Estimation via State Augmentation With Stochastic Permutations

Authors: Edwin V. Bonilla, Pantelis Elinas, He Zhao, Maurizio Filippone, Vassili Kitsios, Terry O'Kane

Abstract: Estimating the structure of a Bayesian network, in the form of a directed acyclic graph (DAG), from observational data is a statistically and computationally hard problem with essential applications in areas such as causal discovery. Bayesian approaches are a promising direction for solving this task, as they allow for uncertainty quantification and deal with well-known identifiability issues. Fro… ▽ More Estimating the structure of a Bayesian network, in the form of a directed acyclic graph (DAG), from observational data is a statistically and computationally hard problem with essential applications in areas such as causal discovery. Bayesian approaches are a promising direction for solving this task, as they allow for uncertainty quantification and deal with well-known identifiability issues. From a probabilistic inference perspective, the main challenges are (i) representing distributions over graphs that satisfy the DAG constraint and (ii) estimating a posterior over the underlying combinatorial space. We propose an approach that addresses these challenges by formulating a joint distribution on an augmented space of DAGs and permutations. We carry out posterior estimation via variational inference, where we exploit continuous relaxations of discrete distributions. We show that our approach performs competitively when compared with a wide range of Bayesian and non-Bayesian benchmarks on a range of synthetic and real datasets. △ Less

Submitted 28 May, 2024; v1 submitted 4 February, 2024; originally announced February 2024.

Realizable Eddy Damped Markovian Anisotropic Closure for Turbulence and Rossby Wave Interactions

Authors: Jorgen S Frederiksen, Terence J O'Kane

Abstract: A realizable Eddy Damped Markovian Anisotropic Closure (EDMAC) is presented for the interaction of two dimensional turbulence and transient waves such as Rossby waves. The structure of the EDMAC ensures that it is as computationally efficient as the Eddy Damped Quasi Normal Markovian (EDQNM) closure but unlike the EDQNM is guaranteed to be realizable in the presence of transient waves. Jack Herrin… ▽ More A realizable Eddy Damped Markovian Anisotropic Closure (EDMAC) is presented for the interaction of two dimensional turbulence and transient waves such as Rossby waves. The structure of the EDMAC ensures that it is as computationally efficient as the Eddy Damped Quasi Normal Markovian (EDQNM) closure but unlike the EDQNM is guaranteed to be realizable in the presence of transient waves. Jack Herring's important contributions to laying the foundations of statistical dynamical closure theories of fluid turbulence are briefly reviewed. The topics covered include equilibrium statistical mechanics, Eulerian and Lagrangian statistical dynamical closure theories, and the statistical dynamics of the interaction of turbulence with topography. The impact of Herring's work is described and placed in the context of related developments. Some of the further works that have built on Herring's foundations are discussed. The relationships between theoretical approaches employed in statistical classical and quantum field theories, and their overlap, are outlined. The seminal advances made by the pioneers in strong interaction fluid turbulence are put into perspective by comparing related developments in strong interaction quantum filed theory. △ Less

Submitted 12 June, 2023; originally announced June 2023.

Comments: 23 pages, 0 figures

Free-Form Variational Inference for Gaussian Process State-Space Models

Authors: Xuhui Fan, Edwin V. Bonilla, Terence J. O'Kane, Scott A. Sisson

Abstract: Gaussian process state-space models (GPSSMs) provide a principled and flexible approach to modeling the dynamics of a latent state, which is observed at discrete-time points via a likelihood model. However, inference in GPSSMs is computationally and statistically challenging due to the large number of latent variables in the model and the strong temporal dependencies between them. In this paper, w… ▽ More Gaussian process state-space models (GPSSMs) provide a principled and flexible approach to modeling the dynamics of a latent state, which is observed at discrete-time points via a likelihood model. However, inference in GPSSMs is computationally and statistically challenging due to the large number of latent variables in the model and the strong temporal dependencies between them. In this paper, we propose a new method for inference in Bayesian GPSSMs, which overcomes the drawbacks of previous approaches, namely over-simplified assumptions, and high computational requirements. Our method is based on free-form variational inference via stochastic gradient Hamiltonian Monte Carlo within the inducing-variable formalism. Furthermore, by exploiting our proposed variational distribution, we provide a collapsed extension of our method where the inducing variables are marginalized analytically. We also showcase results when combining our framework with particle MCMC methods. We show that, on six real-world datasets, our approach can learn transition dynamics and latent states more accurately than competing methods. △ Less

Submitted 16 July, 2023; v1 submitted 20 February, 2023; originally announced February 2023.

Comments: Updating to final version to appear in the proceedings

Physically-inspired computational tools for sharp detection of material inhomogeneities in magnetic imaging

Authors: Illia Horenko, Davi Rodrigues, Terence O'Kane, Karin Everschor-Sitte

Abstract: Detection of material inhomogeneities is an important task in magnetic imaging and plays a significant role in understanding physical processes. For example, in spintronics, the sample heterogeneity determines the onset of current-driven magnetization motion. While often a significant effort is made in enhancing the resolution of an experimental technique to obtain a deeper insight into the physic… ▽ More Detection of material inhomogeneities is an important task in magnetic imaging and plays a significant role in understanding physical processes. For example, in spintronics, the sample heterogeneity determines the onset of current-driven magnetization motion. While often a significant effort is made in enhancing the resolution of an experimental technique to obtain a deeper insight into the physical properties, here we want to emphasize that an advantageous data analysis has the potential to provide a lot more insight into given data set, in particular when being close to the resolution limit where the noise becomes at least of the same order as the signal. In this work, we introduce two tools - the average latent dimension and average latent entropy - which allow for the detection of very subtle material inhomogeneity patterns in the data. For example, for the Ising model, we show that these tools are able to resolve exchange differences down to $1\%$. For a micromagnetic model, we demonstrate that the latent entropy can be used to detect changes in the easy axis anisotropy from magnetization data. We show that the latent entropy remains robust when imposing noise on the data, changing less than $0.3\%$ after adding Gaussian noise of the same amplitude as the signal. Furthermore, we demonstrate that these data-driven tools can be used to visualize inhomogeneities based on MOKE data of magnetic whirls and thereby can help to explicitly resolve impurities and pinning centers. To evaluate the performance of the average latent dimension and entropy, we show that they outperform common instruments ranging from standard statistics measures to state-of-the-art data analysis techniques such as Gaussian mixture models not only in recognition quality but also in the required computational cost. △ Less

Submitted 13 September, 2020; v1 submitted 10 July, 2019; originally announced July 2019.

Comments: 16 pages, 6 figures

Journal ref: Commun. Appl. Math. Comput. Sci. 16 (2021) 267-297

Low-frequency regime transitions and predictability of regimes in a barotropic model

Authors: B. T. Nadiga, T. J. O'Kane

Abstract: Predictability of flow is examined in a barotropic vorticity model that admits low frequency regime transitions between zonal and dipolar states. Such transitions in the model were first studied by Bouchet and Simonnet (2009) and are reminiscent of regime change phenomena in the weather and climate systems wherein extreme and abrupt qualitative changes occur, seemingly randomly, after long periods… ▽ More Predictability of flow is examined in a barotropic vorticity model that admits low frequency regime transitions between zonal and dipolar states. Such transitions in the model were first studied by Bouchet and Simonnet (2009) and are reminiscent of regime change phenomena in the weather and climate systems wherein extreme and abrupt qualitative changes occur, seemingly randomly, after long periods of apparent stability. Mechanisms underlying regime transitions in the model are not well understood yet. From the point of view of atmospheric and oceanic dynamics, a novel aspect of the model is the lack of any source of background gradient of potential-vorticity such as topography or planetary gradient of rotation rate (e.g., as in Charney & DeVore '79). We consider perturbations that are embedded onto the system's chaotic attractor under the full nonlinear dynamics as bred vectors---nonlinear generalizations of the leading (backward) Lyapunov vector. We find that ensemble predictions that use bred vector perturbations are more robust in terms of error-spread relationship than those that use Lyapunov vector perturbations. In particular, when bred vector perturbations are used in conjunction with a simple data assimilation scheme (nudging to truth), we find that at least some of the evolved perturbations align to identify low-dimensional subspaces associated with regions of large forecast error in the control (unperturbed, data-assimilating) run; this happens less often in ensemble predictions that use Lyapunov vector perturbations. Nevertheless, in the inertial regime we consider, we find that (a) the system is more predictable when it is in the zonal regime, and that (b) the horizon of predictability is far too short compared to characteristic time scales associated with processes that lead to regime transitions, thus precluding the possibility of predicting such transitions. △ Less

Submitted 19 May, 2016; originally announced May 2016.

Stochastic Climate Theory and Modelling

Authors: Christian L. E. Franzke, Terence J. O'Kane, Judith Berner, Paul D. Williams, Valerio Lucarini

Abstract: Stochastic methods are a crucial area in contemporary climate research and are increasingly being used in comprehensive weather and climate prediction models as well as reduced order climate models. Stochastic methods are used as subgrid-scale parameterizations as well as for model error representation, uncertainty quantification, data assimilation and ensemble prediction. The need to use stochast… ▽ More Stochastic methods are a crucial area in contemporary climate research and are increasingly being used in comprehensive weather and climate prediction models as well as reduced order climate models. Stochastic methods are used as subgrid-scale parameterizations as well as for model error representation, uncertainty quantification, data assimilation and ensemble prediction. The need to use stochastic approaches in weather and climate models arises because we still cannot resolve all necessary processes and scales in comprehensive numerical weather and climate prediction models. In many practical applications one is mainly interested in the largest and potentially predictable scales and not necessarily in the small and fast scales. For instance, reduced order models can simulate and predict large scale modes. Statistical mechanics and dynamical systems theory suggest that in reduced order models the impact of unresolved degrees of freedom can be represented by suitable combinations of deterministic and stochastic components and non-Markovian (memory) terms. Stochastic approaches in numerical weather and climate prediction models also lead to the reduction of model biases. Hence, there is a clear need for systematic stochastic approaches in weather and climate modelling. In this review we present evidence for stochastic effects in laboratory experiments. Then we provide an overview of stochastic climate theory from an applied mathematics perspectives. We also survey the current use of stochastic methods in comprehensive weather and climate prediction models and show that stochastic parameterizations have the potential to remedy many of the current biases in these comprehensive models. △ Less

Submitted 1 September, 2014; originally announced September 2014.