Gaussian processes for regression

The Bayesian analysis of neural networks is difficult because a simple prior over weights implies a complex prior distribution over functions. In this paper we investigate the use of Gaussian process priors over functions, which permit the predictive Bayesian analysis for fixed values of hyperparameters to be carried out exactly using matrix operations. Two methods, using optimization and averaging (via Hybrid Monte Carlo) over hyperparameters have been tested on a number of challenging problems and have produced excellent results.

Bayesian analysis of the scatterometer wind retrieval inverse problem:Some new approaches

The retrieval of wind vectors from satellite scatterometer observations is a non-linear inverse problem. A common approach to solving inverse problems is to adopt a Bayesian framework and to infer the posterior distribution of the parameters of interest given the observations by using a likelihood model relating the observations to the parameters, and a prior distribution over the parameters. We show how Gaussian process priors can be used efficiently with a variety of likelihood models, using local forward (observation) models and direct inverse models for the scatterometer. We present an enhanced Markov chain Monte Carlo method to sample from the resulting multimodal posterior distribution. We go on to show how the computational complexity of the inference can be controlled by using a sparse, sequential Bayes algorithm for estimation with Gaussian processes. This helps to overcome the most serious barrier to the use of probabilistic, Gaussian process methods in remote sensing inverse problems, which is the prohibitively large size of the data sets. We contrast the sampling results with the approximations that are found by using the sparse, sequential Bayes algorithm.

A pragmatic Bayesian approach to wind field retrieval

The ERS-1 Satellite was launched in July 1991 by the European Space Agency into a polar orbit at about 800 km, carrying a C-band scatterometer. A scatterometer measures the amount of backscatter microwave radiation reflected by small ripples on the ocean surface induced by sea-surface winds, and so provides instantaneous snap-shots of wind flow over large areas of the ocean surface, known as wind fields. Inherent in the physics of the observation process is an ambiguity in wind direction; the scatterometer cannot distinguish if the wind is blowing toward or away from the sensor device. This ambiguity implies that there is a one-to-many mapping between scatterometer data and wind direction. Current operational methods for wind field retrieval are based on the retrieval of wind vectors from satellite scatterometer data, followed by a disambiguation and filtering process that is reliant on numerical weather prediction models. The wind vectors are retrieved by the local inversion of a forward model, mapping scatterometer observations to wind vectors, and minimising a cost function in scatterometer measurement space. This thesis applies a pragmatic Bayesian solution to the problem. The likelihood is a combination of conditional probability distributions for the local wind vectors given the scatterometer data. The prior distribution is a vector Gaussian process that provides the geophysical consistency for the wind field. The wind vectors are retrieved directly from the scatterometer data by using mixture density networks, a principled method to model multi-modal conditional probability density functions. The complexity of the mapping and the structure of the conditional probability density function are investigated. A hybrid mixture density network, that incorporates the knowledge that the conditional probability distribution of the observation process is predominantly bi-modal, is developed. The optimal model, which generalises across a swathe of scatterometer readings, is better on key performance measures than the current operational model. Wind field retrieval is approached from three perspectives. The first is a non-autonomous method that confirms the validity of the model by retrieving the correct wind field 99% of the time from a test set of 575 wind fields. The second technique takes the maximum a posteriori probability wind field retrieved from the posterior distribution as the prediction. For the third technique, Markov Chain Monte Carlo (MCMC) techniques were employed to estimate the mass associated with significant modes of the posterior distribution, and make predictions based on the mode with the greatest mass associated with it. General methods for sampling from multi-modal distributions were benchmarked against a specific MCMC transition kernel designed for this problem. It was shown that the general methods were unsuitable for this application due to computational expense. On a test set of 100 wind fields the MAP estimate correctly retrieved 72 wind fields, whilst the sampling method correctly retrieved 73 wind fields.

Block GTM: Incorporating prior knowledge of covariance structure in data visualisation

Visualising data for exploratory analysis is a big challenge in scientific and engineering domains where there is a need to gain insight into the structure and distribution of the data. Typically, visualisation methods like principal component analysis and multi-dimensional scaling are used, but it is difficult to incorporate prior knowledge about structure of the data into the analysis. In this technical report we discuss a complementary approach based on an extension of a well known non-linear probabilistic model, the Generative Topographic Mapping. We show that by including prior information of the covariance structure into the model, we are able to improve both the data visualisation and the model fit.

Further studies on flow distribution on distillation sieve trays

The thesis describes experimental work on the possibility of using deflection baffles in conventional distillation trays as flow straightening devices, with the view of enhancing tray efficiency. The mode of operation is based on deflecting part of the liquid momentum from the centre of the tray to the segment regions in order to drive stagnating liquid at the edges forward. The first part of the work was a detailed investigation into the two-phase flow patterns produced on a conventional sieve tray having 1 mm hole size perforations. The data provide a check on some earlier work and extend the range of the existing databank, particularly to conditions more typical of industrial operation. A critical survey of data collected on trays with different hole sizes (Hine, 1990; Chambers, 1993; Fenwick, 1996; this work) showed that the hole diameter has a significant influence on the flow regime, the size of the stagnant regions and the hydraulic and mass transfer performance. Five modified tray topologies were created with different configurations of baffles and tested extensively in the 2.44 m diameter air-water pilot distillation simulator for their efficacy in achieving uniform flow across the tray and for their impact on tray loading capacity and mass transfer efficiency. Special attention was given to the calibration of the over 100 temperature probes used in measuring the water temperature across the tray on which the heat and mass transfer analogy is based. In addition to normal tray capacity experiments, higher weir load experiments were conducted using a 'half-tray' mode in order to extend the range of data to conditions more typical of industrial operation. The modified trays show superior flow characteristics compared to the conventional tray in terms of the ability to replenish the zones of exceptionally low temperatures and high residence times at the edges of the tray, to lower the bulk liquid gradient and to achieve a more uniform flow across the tray. These superior flow abilities, however, tend to diminish with increasing weir load because of the increasing tendency for the liquid to jump over the barriers instead of flowing over them. The modified tray topologies showed no tendency to cause undue limitation to tray loading capacity. Although the improvement in the efficiency of a single tray over that of the conventional tray was moderate and in some cases marginal, the multiplier effect in a multiple tray column situation would be significant (Porter et al., 1972). These results are in good agreement with an associated CFD studies (Fischer, 1999) carried out by partners in the Advanced Studies in Distillation consortium. It is concluded that deflection baffles can be used in a conventional distillation sieve tray to achieve better liquid flow distribution and obtain enhanced mass transfer efficiency, without undermining the tray loading capacity. Unlike any other controlled-flow tray whose mechanical complexity impose stringent manufacturing and installation tolerances, the baffled-tray models are simple to design, manufacture and install and thus provide an economic method of retrofitting badly performing sieve trays both in terms of downtime and fabrication. NOTE APPENDICES 2-5 ARE ON A SEPARATE FLOPPY DISK ONLY AVAILABLE FOR CONSULTATION AT ASTON UNIVERSITY LIBRARY WITH PRIOR ARRANGEMENT

Data visualisation and exploration with prior knowledge

Visualising data for exploratory analysis is a major challenge in many applications. Visualisation allows scientists to gain insight into the structure and distribution of the data, for example finding common patterns and relationships between samples as well as variables. Typically, visualisation methods like principal component analysis and multi-dimensional scaling are employed. These methods are favoured because of their simplicity, but they cannot cope with missing data and it is difficult to incorporate prior knowledge about properties of the variable space into the analysis; this is particularly important in the high-dimensional, sparse datasets typical in geochemistry. In this paper we show how to utilise a block-structured correlation matrix using a modification of a well known non-linear probabilistic visualisation model, the Generative Topographic Mapping (GTM), which can cope with missing data. The block structure supports direct modelling of strongly correlated variables. We show that including prior structural information it is possible to improve both the data visualisation and the model fit. These benefits are demonstrated on artificial data as well as a real geochemical dataset used for oil exploration, where the proposed modifications improved the missing data imputation results by 3 to 13%.

Toxicity as a factor influencing the distribution of gammorus pulex in some Midland rivers

DUE TO COPYRIGHT RESTRICTIONS ONLY AVAILABLE FOR CONSULTATION AT ASTON UNIVERSITY LIBRARY AND INFORMATION SERVICES WITH PRIOR ARRANGEMENT

The distribution and paragenesis of low temperature nickel/iron/cobalt sulphides in central Britain

DUE TO COPYRIGHT RESTRICTIONS ONLY AVAILABLE FOR CONSULTATION AT ASTON UNIVERSITY LIBRARY AND INFORMATION SERVICES WITH PRIOR ARRANGEMENT

Incorporating sentiment prior knowledge for weakly supervised sentiment analysis

This article presents two novel approaches for incorporating sentiment prior knowledge into the topic model for weakly supervised sentiment analysis where sentiment labels are considered as topics. One is by modifying the Dirichlet prior for topic-word distribution (LDA-DP), the other is by augmenting the model objective function through adding terms that express preferences on expectations of sentiment labels of the lexicon words using generalized expectation criteria (LDA-GE). We conducted extensive experiments on English movie review data and multi-domain sentiment dataset as well as Chinese product reviews about mobile phones, digital cameras, MP3 players, and monitors. The results show that while both LDA-DP and LDAGE perform comparably to existing weakly supervised sentiment classification algorithms, they are much simpler and computationally efficient, rendering themmore suitable for online and real-time sentiment classification on the Web. We observed that LDA-GE is more effective than LDA-DP, suggesting that it should be preferred when considering employing the topic model for sentiment analysis. Moreover, both models are able to extract highly domain-salient polarity words from text.

Proposals for more effective distribution procedures

DUE TO COPYRIGHT RESTRICTIONS ONLY AVAILABLE FOR CONSULTATION AT ASTON UNIVERSITY LIBRARY AND INFORMATION SERVICES WITH PRIOR ARRANGEMENT

Investigation into bilateral distribution of cerebral blood flow using a multidetector scintillating system

DUE TO COPYRIGHT RESTRICTIONS ONLY AVAILABLE FOR CONSULTATION AT ASTON UNIVERSITY LIBRARY AND INFORMATION SERVICES WITH PRIOR ARRANGEMENT

The use of an energy analyser for the measurement of the total energy distribution of field-emitted electrons from tungsten

DUE TO COPYRIGHT RESTRICTIONS ONLY AVAILABLE FOR CONSULTATION AT ASTON UNIVERSITY LIBRARY AND INFORMATION SERVICES WITH PRIOR ARRANGEMENT

A simulation analysis of spoke-terminals operating in LTL Hub-and-Spoke freight distribution systems

DUE TO COPYRIGHT RESTRICTIONS ONLY AVAILABLE FOR CONSULTATION AT ASTON UNIVERSITY LIBRARY AND INFORMATION SERVICES WITH PRIOR ARRANGEMENT The research presented in this thesis is concerned with Discrete-Event Simulation (DES) modelling as a method to facilitate logistical policy development within the UK Less-than-Truckload (LTL) freight distribution sector which has been typified by “Pallet Networks” operating on a hub-and-spoke philosophy. Current literature relating to LTL hub-and-spoke and cross-dock freight distribution systems traditionally examines a variety of network and hub design configurations. Each is consistent with classical notions of creating process efficiency, improving productivity, reducing costs and generally creating economies of scale through notions of bulk optimisation. Whilst there is a growing abundance of papers discussing both the network design and hub operational components mentioned above, there is a shortcoming in the overall analysis when it comes to discussing the “spoke-terminal” of hub-and-spoke freight distribution systems and their capabilities for handling the diverse and discrete customer profiles of freight that multi-user LTL hub-and-spoke networks typically handle over the “last-mile” of the delivery, in particular, a mix of retail and non-retail customers. A simulation study is undertaken to investigate the impact on operational performance when the current combined spoke-terminal delivery tours are separated by ‘profile-type’ (i.e. retail or nonretail). The results indicate that a potential improvement in delivery performance can be made by separating retail and non-retail delivery runs at the spoke-terminal and that dedicated retail and non-retail delivery tours could be adopted in order to improve customer delivery requirements and adapt hub-deployed policies. The study also leverages key operator experiences to highlight the main practical implementation challenges when integrating the observed simulation results into the real-world. The study concludes that DES be harnessed as an enabling device to develop a ‘guide policy’. This policy needs to be flexible and should be applied in stages, taking into account the growing retail-exposure.