Search

This thesis is a biographical study of John of Brienne, king of Jerusalem and later Latin emperor of Constantinople (d. 1237). John’s extraordinary career is touched on by many commentators concerned with the crusades and the Latin East in the early thirteenth century, but it has not been properly re-assessed for more than seventy years. A comprehensive re-examination opens up new angles on the political structures and social landscapes that produced it. John’s career illustrates some residual strengths of the Jerusalemite monarchy just before the start of the Hohenstaufen epoch. It also sheds light on a period in the history of the Latin empire all too easily regarded as largely a void. But within the biographical context, the thesis’s focus is more on the complex interplay between the Latin West and East in the early thirteenth century. A principal theme in this regard is the mobility, in geographical and politico-hierarchical terms, of a specific echelon of the high aristocracy in early thirteenth-century Europe, building on Bartlett’s conception of the contemporaneous western European ‘aristocratic diaspora’. Aristocrats who are ‘not quite first rank’ can be discerned on the make in regions, both west and east, distant from their original homelands. Much of the significance of that lies in the context, the variety of opportunities, and also the limitations on such figures. Whilst this thesis dwells on John’s experience of patronage and dependency, it also identifies grounds for tensions in his ‘new’ environments, as well as highlighting the opportunities and pitfalls presented by ‘dynastic interstices’. In this way, the thesis unpacks many of the ‘more normal’ features of the aristocratic diaspora out of John’s exceptional career. The thesis links together the thematic material to focus, in particular, on the interactions between various Western great powers and John as a client figure.

A methodology is developed to estimate comparable international migration flows between a set of countries. International migration flow data may be missing, reported by the sending country, reported by the receiving country or reported by both the sending and receiving countries. For the last situation, reported counts rarely match due to differences in defnitions and data collection systems. In this thesis, reported counts are harmonized using correction factors estimated from a constrained optimization procedure. Factors are applied to scale data known to be of a reliable standard, creating an incomplete migration flow table of harmonized values. Cells for which no reliable reported flows exist are then estimated from a negative binomial regression model fitted using the Expectation- Maximization (EM) type algorithm. Covariate information for this model is drawn from international migration theory. Finally, measures of precision for all missing cell estimates are derived using the Supplemented EM algorithm. Recent data on international migration between countries in Europe are used to illustrate the methodology. The results represent a complete table of comparable ows that can be used by regional policy makers and social scientist alike to better understand population behaviour and change.

The present work falls naturally into several distinct parts. In the first place the history of the movement has been described in some detail. It was found difficult to confine the history to the development of the clinic only as, after all, the clinic is only part of a larger scheme. The historical portion therefore treats of the development of the entire Maternity and Child Welfare movement from the earliest times. Some apology is perhaps required if the reader finds the historical portion lengthy but the material has been so generous and interesting that it was felt a shorter sketch would not do the matter justice. During the earlier portions of the history it was convenient to describe both the Maternity and Infant parts of the movement together; but with the advent of the twentieth century the development of the two parts became so rapid and involved that it was more convenient to describe from that point each part of the movement separately. In the succeeding portions of the work the Infant Welfare and Ante Natal Clinic are discussed. The functions, future development and value of the clinic are considered in somedetail as well as the shortcomings of the present system. The remaining portions of the work are concerned with a short summary in which the principal findings are embodied. A bibliography of the various books and papers consulted or read is appended.

Sleep-disordered breathing (SDB) is a serious and prevalent condition that results from the collapse of the upper airway during sleep, which leads to oxygen desaturations, unphysiological variations in intrathoracic pressure, and sleep fragmentation. Its most common form is obstructive sleep apnoea (OSA). This has a big impact on quality of life, and is associated with cardiovascular morbidity. Polysomnography, the gold standard for diagnosing SDB, is obtrusive, time-consuming and expensive. Alternative diagnostic approaches have been proposed to overcome its limitations. In particular, acoustic analysis of sleep breathing sounds offers an unobtrusive and inexpensive means to screen for SDB, since it displays symptoms with unique acoustic characteristics. These include snoring, loud gasps, chokes, and absence of breathing. This thesis investigates deep learning methods, which have revolutionised speech and audio technology, to robustly screen for SDB in typical sleep conditions using acoustics. To begin with, the desirable characteristics for an acoustic corpus of SDB, and the acoustic definition of snoring are considered to create corpora for this study. Then three approaches are developed to tackle increasingly complex scenarios. Firstly, with the aim of leveraging a large amount of unlabelled SDB data, unsupervised learning is applied to learn novel feature representations with deep neural networks for the classification of SDB events such as snoring. The incorporation of contextual information to assist the classifier in producing realistic event durations is investigated. Secondly, the temporal pattern of sleep breathing sounds is exploited using convolutional neural networks to screen participants sleeping by themselves for OSA. The integration of acoustic features with physiological data for screening is examined. Thirdly, for the purpose of achieving robustness to bed partner breathing sounds, recurrent neural networks are used to screen a subject and their bed partner for SDB in the same session. Experiments conducted on the constructed corpora show that the developed systems accurately classify SDB events, screen for OSA with high sensitivity and specificity, and screen a subject and their bed partner for SDB with encouraging performance. In conclusion, this thesis makes promising progress in improving access to SDB diagnosis through low-cost and non-invasive methods.

Users have access to a diverse set of interfaces that can be used to interact with software. Tools exist for automatically generating test data for an application, but the data required by each user interface is complex. Generating realistic data similar to that of a user is difficult. The environment which an application is running inside may also limit the data available, or updates to an operating system can break support for tools that generate test data. Consequently, applications exist for which there are no automated methods of generating test data similar to that which a user would provide through real usage of a user interface. With no automated method of generating data, the cost of testing increases and there is an increased chance of bugs being released into production code. In this thesis, we investigate techniques which aim to mimic users, observing how stored user interactions can be split to generate data targeted at specific states of an application, or to generate different subareas of the data structure provided by a user interface. To reduce the cost of gathering and labelling graphical user interface data, we look at generating randomised screen shots of applications, which can be automatically labelled and used in the training stage of a machine learning model. These trained models could guide a randomised approach at generating tests, achieving a significantly higher branch coverage than an unguided random approach. However, for natural user interfaces, which allow interaction through body tracking, we could not learn such a model through generated data. We find that models derived from real user data can generate tests with a significantly higher branch coverage than a purely random tester for both natural and graphical user interfaces. Our approaches use no feedback from an application during test generation. Consequently, the models are “generating data in the dark”. Despite this, these models can still generate tests with a higher coverage than random testing, but there may be a benefit to inferring the current state of an application and using this to guide data generation.

In an effort to understand intrinsic grain boundary brittleness, computer simulation of grain boundary structures in ordered NiAl was carried out. The considered boundaries were boundaries of low reciprocal coincident site density Σ. The structural unit model was used to obtain a general picture of the grain boundary features in NiAl. Relation between the grain boundary stoichiometry and the energy was studied and it was found that the grain boundary energy increases with a higher aluminum content in the grain boundary. Thus, excess of aluminum in the grain boundary seems to make the material brittle. The grain boundaries obtained are relatively dense and grain boundary brittleness would not be expected to originate from large interstitial holes. However, the structures computed show a large spread in energy values and it is proposed that the observed brittleness is due to the appearance of high energy structures when dislocations arrive at the boundary.

This thesis investigates methods for visual speech enhancement to support auditory and audiovisual speech perception. Normal-hearing non-native listeners receiving cochlear implant (CI) simulated speech are used as ‘proxy’ listeners for CI users, a proposed user group who could benefit from such enhancement methods in speech perception training. Both CI users and non-native listeners share similarities with regards to audiovisual speech perception, including increased sensitivity to visual speech cues. Two enhancement methods are proposed: (i) an appearance based method, which modifies the appearance of a talker’s lips using colour and luminance blending to apply a ‘lipstick effect’ to increase the saliency of mouth shapes; and (ii) a kinematics based method, which amplifies the kinematics of the talker’s mouth to create the effect of more pronounced speech (an ‘exaggeration effect’). The application that is used to test the enhancements is speech perception training, or audiovisual training, which can be used to improve listening skills. An audiovisual training framework is presented which structures the evaluation of the effectiveness of these methods. It is used in two studies. The first study, which evaluates the effectiveness of the lipstick effect, found a significant improvement in audiovisual and auditory perception. The second study, which evaluates the effectiveness of the exaggeration effect, found improvement in the audiovisual perception of a number of phoneme classes; no evidence was found of improvements in the subsequent auditory perception, as audiovisual recalibration to visually exaggerated speech may have impeded learning when used in the audiovisual training. The thesis also investigates an example of kinematics based enhancement which is observed in Lombard speech, by studying the behaviour of visual Lombard phonemes in different contexts. Due to the lack of suitable datasets for this analysis, the thesis presents a novel audiovisual Lombard speech dataset recorded under high SNR, which offers two, fixed head-pose, synchronised views of each talker in the dataset.

This thesis explores compensation for reverberation in human listeners and machines. Late reverberation is typically understood as a distortion which degrades intelligibility. Recent research, however, shows that late reverberation is not always detrimental to human speech perception. At times, prolonged exposure to reverberation can provide a helpful acoustic context which improves identification of reverberant speech sounds. The physiology underpinning our robustness to reverberation has not yet been elucidated, but is speculated in this thesis to include efferent processes which have previously been shown to improve discrimination of noisy speech. These efferent pathways descend from higher auditory centres, effectively recalibrating the encoding of sound in the cochlea. Moreover, this thesis proposes that efferent-inspired computational models based on psychoacoustic principles may also improve performance for machine listening systems in reverberant environments. A candidate model for perceptual compensation for reverberation is proposed in which efferent suppression derives from the level of reverberation detected in the simulated auditory nerve response. The model simulates human performance in a phoneme-continuum identification task under a range of reverberant conditions, where a synthetically controlled test-word and its surrounding context phrase are independently reverberated. Addressing questions which arose from the model, a series of perceptual experiments used naturally spoken speech materials to investigate aspects of the psychoacoustic mechanism underpinning compensation. These experiments demonstrate a monaural compensation mechanism that is influenced by both the preceding context (which need not be intelligible speech) and by the test-word itself, and which depends on the time-direction of reverberation. Compensation was shown to act rapidly (within a second or so), indicating a monaural mechanism that is likely to be effective in everyday listening. Finally, the implications of these findings for the future development of computational models of auditory perception are considered.

Understanding the design of talk-in-interaction is important in many domains, including speech technology. Although phonetic, linguistic and gestural correlates have been identified for some of the social actions that conversational participants accomplish, it is only recently that researchers have begun to take account of the immediately prior interactional context as an important factor influencing the design of a speaker’s turn. The present study explores the influence of context by focussing on characteristics of short turns produced by one speaker between turns from another speaker. The hypothesis is that the speaker designs her inserted turn as a match to the prior turn when wishing to align with the previous speaker’s agenda. By contrast, non-matching would display that the speaker is non-aligning, preferring instead to initiate a new action for example. Data are taken from the AMI corpus, focussing on the spontaneous talk of first-language English participants. Using sequential analysis, such short turns are classified as either aligning or non-aligning in accordance with definitions in the Conversation Analysis literature. The degree of prosodic similarity between the inserted turn and the prior speaker’s turn is measured using novel acoustic techniques. The results show that aligning turns are significantly more similar to the immediately preceding turn, in terms of pitch contour, than non-aligning turns. In contrast to the prosodic-acoustic analysis, the results of the gestural analysis indicate that aligning and non-aligning are differentiated by the use of distinct gestures, rather than by the matching (or non-matching) of gestures across the adjacent turns. These results support the view that choice of pitch contour is managed locally, rather than by reference to an intonational lexicon. However, this is not the case for speakers’ use of gesture. The implications of these findings for a model of talk-in-interaction are considered, along with potential applications.