Tractable nonparametric Bayesian inference in Poisson processes with Gaussian process intensities

The inhomogeneous Poisson process is a point process that has varying intensity across its domain (usually time or space). For nonparametric Bayesian modeling, the Gaussian process is a useful way to place a prior distribution on this intensity. The combination of a Poisson process and GP is known as a Gaussian Cox process, or doubly-stochastic Poisson process. Likelihood-based inference in these models requires an intractable integral over an infinite-dimensional random function. In this paper we present the first approach to Gaussian Cox processes in which it is possible to perform inference without introducing approximations or finite-dimensional proxy distributions. We call our method the Sigmoidal Gaussian Cox Process, which uses a generative model for Poisson data to enable tractable inference via Markov chain Monte Carlo. We compare our methods to competing methods on synthetic data and apply it to several real-world data sets.

Markov Switching Risk Premium and the term structure of interest rates. Empirical evidence from US post-war interest rates

This paper considers the basic present value model of interest rates under rational expectations with two additional features. First, following McCallum (1994), the model assumes a policy reaction function where changes in the short-term interest rate are determined by the long-short spread. Second, the short-term interest rate and the risk premium processes are characterized by a Markov regime-switching model. Using US post-war interest rate data, this paper finds evidence that a two-regime switching model fits the data better than the basic model. The estimation results also show the presence of two alternative states displaying quite different features.

Segmentação de nome e endereço por meio de modelos escondidos de Markov e sua aplicação em processos de vinculação de registros

A segmentação dos nomes nas suas partes constitutivas é uma etapa fundamental no processo de integração de bases de dados por meio das técnicas de vinculação de registros. Esta separação dos nomes pode ser realizada de diferentes maneiras. Este estudo teve como objetivo avaliar a utilização do Modelo Escondido de Markov (HMM) na segmentação nomes e endereços de pessoas e a eficiência desta segmentação no processo de vinculação de registros. Foram utilizadas as bases do Sistema de Informações sobre Mortalidade (SIM) e do Subsistema de Informação de Procedimentos de Alta Complexidade (APAC) do estado do Rio de Janeiro no período entre 1999 a 2004. Uma metodologia foi proposta para a segmentação de nome e endereço sendo composta por oito fases, utilizando rotinas implementadas em PL/SQL e a biblioteca JAHMM, implementação na linguagem Java de algoritmos de HMM. Uma amostra aleatória de 100 registros de cada base foi utilizada para verificar a correção do processo de segmentação por meio do modelo HMM.Para verificar o efeito da segmentação do nome por meio do HMM, três processos de vinculação foram aplicados sobre uma amostra das duas bases citadas acima, cada um deles utilizando diferentes estratégias de segmentação, a saber: 1) divisão dos nomes pela primeira parte, última parte e iniciais do nome do meio; 2) divisão do nome em cinco partes; (3) segmentação segundo o HMM. A aplicação do modelo HMM como mecanismo de segmentação obteve boa concordância quando comparado com o observador humano. As diferentes estratégias de segmentação geraram resultados bastante similares na vinculação de registros, tendo a estratégia 1 obtido um desempenho pouco melhor que as demais. Este estudo sugere que a segmentação de nomes brasileiros por meio do modelo escondido de Markov não é mais eficaz do que métodos tradicionais de segmentação.

Tree-Structured Stick Breaking Processes for Hierarchical Data

Many data are naturally modeled by an unobserved hierarchical structure. In this paper we propose a flexible nonparametric prior over unknown data hierarchies. The approach uses nested stick-breaking processes to allow for trees of unbounded width and depth, where data can live at any node and are infinitely exchangeable. One can view our model as providing infinite mixtures where the components have a dependency structure corresponding to an evolutionary diffusion down a tree. By using a stick-breaking approach, we can apply Markov chain Monte Carlo methods based on slice sampling to perform Bayesian inference and simulate from the posterior distribution on trees. We apply our method to hierarchical clustering of images and topic modeling of text data.

Copula Processes

We define a copula process which describes the dependencies between arbitrarily many random variables independently of their marginal distributions. As an example, we develop a stochastic volatility model, Gaussian Copula Process Volatility (GCPV), to predict the latent standard deviations of a sequence of random variables. To make predictions we use Bayesian inference, with the Laplace approximation, and with Markov chain Monte Carlo as an alternative. We find both methods comparable. We also find our model can outperform GARCH on simulated and financial data. And unlike GARCH, GCPV can easily handle missing data, incorporate covariates other than time, and model a rich class of covariance structures.

Enhanced poisson sum representation for alpha-stable processes

In this paper we present Poisson sum series representations for α-stable (αS) random variables and a-stable processes, in particular concentrating on continuous-time autoregressive (CAR) models driven by α-stable Lévy processes. Our representations aim to provide a conditionally Gaussian framework, which will allow parameter estimation using Rao-Blackwellised versions of state of the art Bayesian computational methods such as particle filters and Markov chain Monte Carlo (MCMC). To overcome the issues due to truncation of the series, novel residual approximations are developed. Simulations demonstrate the potential of these Poisson sum representations for inference in otherwise intractable α-stable models. © 2011 IEEE.

Gaussian processes for fast policy optimisation of POMDP-based dialogue managers

Modelling dialogue as a Partially Observable Markov Decision Process (POMDP) enables a dialogue policy robust to speech understanding errors to be learnt. However, a major challenge in POMDP policy learning is to maintain tractability, so the use of approximation is inevitable. We propose applying Gaussian Processes in Reinforcement learning of optimal POMDP dialogue policies, in order (1) to make the learning process faster and (2) to obtain an estimate of the uncertainty of the approximation. We first demonstrate the idea on a simple voice mail dialogue task and then apply this method to a real-world tourist information dialogue task. © 2010 Association for Computational Linguistics.

State estimation schemes for independent component coupled hidden Markov models

Conventional Hidden Markov models generally consist of a Markov chain observed through a linear map corrupted by additive noise. This general class of model has enjoyed a huge and diverse range of applications, for example, speech processing, biomedical signal processing and more recently quantitative finance. However, a lesser known extension of this general class of model is the so-called Factorial Hidden Markov Model (FHMM). FHMMs also have diverse applications, notably in machine learning, artificial intelligence and speech recognition [13, 17]. FHMMs extend the usual class of HMMs, by supposing the partially observed state process is a finite collection of distinct Markov chains, either statistically independent or dependent. There is also considerable current activity in applying collections of partially observed Markov chains to complex action recognition problems, see, for example, [6]. In this article we consider the Maximum Likelihood (ML) parameter estimation problem for FHMMs. Much of the extant literature concerning this problem presents parameter estimation schemes based on full data log-likelihood EM algorithms. This approach can be slow to converge and often imposes heavy demands on computer memory. The latter point is particularly relevant for the class of FHMMs where state space dimensions are relatively large. The contribution in this article is to develop new recursive formulae for a filter-based EM algorithm that can be implemented online. Our new formulae are equivalent ML estimators, however, these formulae are purely recursive and so, significantly reduce numerical complexity and memory requirements. A computer simulation is included to demonstrate the performance of our results. © Taylor & Francis Group, LLC.

Bayesian Structured Prediction Using Gaussian Processes

We introduce a conceptually novel structured prediction model, GPstruct, which is kernelized, non-parametric and Bayesian, by design. We motivate the model with respect to existing approaches, among others, conditional random fields (CRFs), maximum margin Markov networks (M3N), and structured support vector machines (SVMstruct), which embody only a subset of its properties. We present an inference procedure based on Markov Chain Monte Carlo. The framework can be instantiated for a wide range of structured objects such as linear chains, trees, grids, and other general graphs. As a proof of concept, the model is benchmarked on several natural language processing tasks and a video gesture segmentation task involving a linear chain structure. We show prediction accuracies for GPstruct which are comparable to or exceeding those of CRFs and SVMstruct.

Gaussian processes for POMDP-based dialogue manager optimization

A partially observable Markov decision process (POMDP) has been proposed as a dialog model that enables automatic optimization of the dialog policy and provides robustness to speech understanding errors. Various approximations allow such a model to be used for building real-world dialog systems. However, they require a large number of dialogs to train the dialog policy and hence they typically rely on the availability of a user simulator. They also require significant designer effort to hand-craft the policy representation. We investigate the use of Gaussian processes (GPs) in policy modeling to overcome these problems. We show that GP policy optimization can be implemented for a real world POMDP dialog manager, and in particular: 1) we examine different formulations of a GP policy to minimize variability in the learning process; 2) we find that the use of GP increases the learning rate by an order of magnitude thereby allowing learning by direct interaction with human users; and 3) we demonstrate that designer effort can be substantially reduced by basing the policy directly on the full belief space thereby avoiding ad hoc feature space modeling. Overall, the GP approach represents an important step forward towards fully automatic dialog policy optimization in real world systems. © 2013 IEEE.

Classification using log Gaussian Cox processes

McCullagh and Yang (2006) suggest a family of classification algorithms based on Cox processes. We further investigate the log Gaussian variant which has a number of appealing properties. Conditioned on the covariates, the distribution over labels is given by a type of conditional Markov random field. In the supervised case, computation of the predictive probability of a single test point scales linearly with the number of training points and the multiclass generalization is straightforward. We show new links between the supervised method and classical nonparametric methods. We give a detailed analysis of the pairwise graph representable Markov random field, which we use to extend the model to semi-supervised learning problems, and propose an inference method based on graph min-cuts. We give the first experimental analysis on supervised and semi-supervised datasets and show good empirical performance.

Equivalence and Reduction of Hidden Markov Models

This report studies when and why two Hidden Markov Models (HMMs) may represent the same stochastic process. HMMs are characterized in terms of equivalence classes whose elements represent identical stochastic processes. This characterization yields polynomial time algorithms to detect equivalent HMMs. We also find fast algorithms to reduce HMMs to essentially unique and minimal canonical representations. The reduction to a canonical form leads to the definition of 'Generalized Markov Models' which are essentially HMMs without the positivity constraint on their parameters. We discuss how this generalization can yield more parsimonious representations of stochastic processes at the cost of the probabilistic interpretation of the model parameters.

Catalyst, additive and substrate effects in intramolecular aromatic additions of alpha-diazoketones

This thesis is focused on transition metal catalysed reaction of α-diazoketones leading to aromatic addition to form azulenones, with particular emphasis on enantiocontrol through use of chiral copper catalysts. The first chapter provides an overview of the influence of variation of the substituent at the diazo carbon on the outcome of subsequent reaction pathways, focusing in particular on C-H insertion, cyclopropanation, aromatic addition and ylide formation drawing together for the first time input from a range of primary reports. Chapter two describes the synthesis of a range of novel α-diazoketones. Rhodium and copper catalysed cyclisation of these to form a range of azulenones is described. Variation of the transition metal catalyst was undertaken using both copper and rhodium based systems and ligand variation, including the design and synthesis of a novel bisoxazoline ligand. The influence of additives, especially NaBARF, on the enantiocontrol was explored in detail and displayed an interesting impact which was sensitive to substituent effects. Further exploration demonstrated that it is the sodium cation which is critical in the additive effects. For the first time, enantiocontrol in the aromatic addition of terminal diazoketones was demonstrated indicating enantiofacial control in the aromatic addition is feasible in the absence of a bridgehead substituent. Determination of the enantiopurity in these compounds was particularly challenging due to the lability of the products. A substantial portion of the work was focused on determining the stereochemical outcome of the aromatic addition processes, both the absolute stereochemistry and extent of enantiopurity. Formation of PTAD adducts was beneficial in this regard. The third chapter contains the full experimental details and spectral characterisation of all novel compounds synthesised in this project, while details of chiral stationary phase HPLC and 1H NMR analysis are included in the appendix.

Latent Stick-Breaking Processes.

We develop a model for stochastic processes with random marginal distributions. Our model relies on a stick-breaking construction for the marginal distribution of the process, and introduces dependence across locations by using a latent Gaussian copula model as the mechanism for selecting the atoms. The resulting latent stick-breaking process (LaSBP) induces a random partition of the index space, with points closer in space having a higher probability of being in the same cluster. We develop an efficient and straightforward Markov chain Monte Carlo (MCMC) algorithm for computation and discuss applications in financial econometrics and ecology. This article has supplementary material online.

Dramatic variability of the carbonate system at a temperate coastal ocean site (Beaufort, North Carolina, USA) is regulated by physical and biogeochemical processes on multiple timescales.

Increasing atmospheric carbon dioxide (CO2) from anthropogenic sources is acidifying marine environments resulting in potentially dramatic consequences for the physical, chemical and biological functioning of these ecosystems. If current trends continue, mean ocean pH is expected to decrease by ~0.2 units over the next ~50 years. Yet, there is also substantial temporal variability in pH and other carbon system parameters in the ocean resulting in regions that already experience change that exceeds long-term projected trends in pH. This points to short-term dynamics as an important layer of complexity on top of long-term trends. Thus, in order to predict future climate change impacts, there is a critical need to characterize the natural range and dynamics of the marine carbonate system and the mechanisms responsible for observed variability. Here, we present pH and dissolved inorganic carbon (DIC) at time intervals spanning 1 hour to >1 year from a dynamic, coastal, temperate marine system (Beaufort Inlet, Beaufort NC USA) to characterize the carbonate system at multiple time scales. Daily and seasonal variation of the carbonate system is largely driven by temperature, alkalinity and the balance between primary production and respiration, but high frequency change (hours to days) is further influenced by water mass movement (e.g. tides) and stochastic events (e.g. storms). Both annual (~0.3 units) and diurnal (~0.1 units) variability in coastal ocean acidity are similar in magnitude to 50 year projections of ocean acidity associated with increasing atmospheric CO2. The environmental variables driving these changes highlight the importance of characterizing the complete carbonate system rather than just pH. Short-term dynamics of ocean carbon parameters may already exert significant pressure on some coastal marine ecosystems with implications for ecology, biogeochemistry and evolution and this shorter term variability layers additive effects and complexity, including extreme values, on top of long-term trends in ocean acidification.