The development and field test of a Mealtime Interaction Clinical Observation Tool: a pilot study and clinical research portfolio

Objective: The purpose of this study was to develop and test psychometric properties of a Mealtime Interaction Clinical Observation Tool (MICOT) that could be used to facilitate assessment and behavioural intervention in childhood feeding difficulties. Methods: Thematic analysis of four focus groups with feeding and behaviour experts identified the content and structure of the MICOT. Following refinement, inter-rater reliability was tested between three healthcare professionals. Results: Six themes were identified for the MICOT, which utilises a traffic-light system to identify areas of strength and areas for intervention. Despite poor inter-rater reliability, for which a number of reasons are postulated, some correlation between psychologists’ ratings was evident. Healthcare professionals liked the tool and reported that it could have good clinical utility. Conclusion: The study provides a promising first version of a clinical observation tool that facilitates assessment and behavioural intervention in childhood feeding difficulties.

Using interaction data for improving the offline and online evaluation of search engines

This thesis investigates how web search evaluation can be improved using historical interaction data. Modern search engines combine offline and online evaluation approaches in a sequence of steps that a tested change needs to pass through to be accepted as an improvement and subsequently deployed. We refer to such a sequence of steps as an evaluation pipeline. In this thesis, we consider the evaluation pipeline to contain three sequential steps: an offline evaluation step, an online evaluation scheduling step, and an online evaluation step. In this thesis we show that historical user interaction data can aid in improving the accuracy or efficiency of each of the steps of the web search evaluation pipeline. As a result of these improvements, the overall efficiency of the entire evaluation pipeline is increased. Firstly, we investigate how user interaction data can be used to build accurate offline evaluation methods for query auto-completion mechanisms. We propose a family of offline evaluation metrics for query auto-completion that represents the effort the user has to spend in order to submit their query. The parameters of our proposed metrics are trained against a set of user interactions recorded in the search engine’s query logs. From our experimental study, we observe that our proposed metrics are significantly more correlated with an online user satisfaction indicator than the metrics proposed in the existing literature. Hence, fewer changes will pass the offline evaluation step to be rejected after the online evaluation step. As a result, this would allow us to achieve a higher efficiency of the entire evaluation pipeline. Secondly, we state the problem of the optimised scheduling of online experiments. We tackle this problem by considering a greedy scheduler that prioritises the evaluation queue according to the predicted likelihood of success of a particular experiment. This predictor is trained on a set of online experiments, and uses a diverse set of features to represent an online experiment. Our study demonstrates that a higher number of successful experiments per unit of time can be achieved by deploying such a scheduler on the second step of the evaluation pipeline. Consequently, we argue that the efficiency of the evaluation pipeline can be increased. Next, to improve the efficiency of the online evaluation step, we propose the Generalised Team Draft interleaving framework. Generalised Team Draft considers both the interleaving policy (how often a particular combination of results is shown) and click scoring (how important each click is) as parameters in a data-driven optimisation of the interleaving sensitivity. Further, Generalised Team Draft is applicable beyond domains with a list-based representation of results, i.e. in domains with a grid-based representation, such as image search. Our study using datasets of interleaving experiments performed both in document and image search domains demonstrates that Generalised Team Draft achieves the highest sensitivity. A higher sensitivity indicates that the interleaving experiments can be deployed for a shorter period of time or use a smaller sample of users. Importantly, Generalised Team Draft optimises the interleaving parameters w.r.t. historical interaction data recorded in the interleaving experiments. Finally, we propose to apply the sequential testing methods to reduce the mean deployment time for the interleaving experiments. We adapt two sequential tests for the interleaving experimentation. We demonstrate that one can achieve a significant decrease in experiment duration by using such sequential testing methods. The highest efficiency is achieved by the sequential tests that adjust their stopping thresholds using historical interaction data recorded in diagnostic experiments. Our further experimental study demonstrates that cumulative gains in the online experimentation efficiency can be achieved by combining the interleaving sensitivity optimisation approaches, including Generalised Team Draft, and the sequential testing approaches. Overall, the central contributions of this thesis are the proposed approaches to improve the accuracy or efficiency of the steps of the evaluation pipeline: the offline evaluation frameworks for the query auto-completion, an approach for the optimised scheduling of online experiments, a general framework for the efficient online interleaving evaluation, and a sequential testing approach for the online search evaluation. The experiments in this thesis are based on massive real-life datasets obtained from Yandex, a leading commercial search engine. These experiments demonstrate the potential of the proposed approaches to improve the efficiency of the evaluation pipeline.

The subjunctive mood in Late Middle English adverbial clauses: the interaction of form and function

This thesis focuses on the history of the inflexional subjunctive and its functional substitutes in Late Middle English. To explore why and how the inflexional subjunctive declined in the history of English language, I analysed 2653 examples of three adverbial clauses introduced by if (1882 examples), though (305 examples) and lest (466 examples). Using a corpus-based approach, this thesis argues that linguistic change in subjunctive constructions did not happen suddenly but rather gradually, and the way it changed was varied , and that different constructions changed at different speeds in different environments. It is well known that the inflexional subjunctive declined in the history of English, mainly because of inflexional loss. Strangely however this topic has been comparatively neglected in the scholarly literature, especially with regard to the Middle English period, probably due to the limitations of data and also because study of this development requires very cumbersome textual research. This thesis has derived and analysed the data from three large corpora in the public domain: the Middle English Grammar Corpus (MEG-C for short), the Innsbruck Computer Archive of Machine-Readable English Texts (ICAMET for short), and some selected texts from The Corpus of Middle English Prose and Verse, part of the Middle English Compendium that also includes the Middle English Dictionary. The data were analysed from three perspectives: 1) clausal type, 2) dialect, and 3) textual genre. The basic methodology for the research was to analyse the examples one by one, with special attention being paid to the peculiarities of each text. In addition, this thesis draw on some complementary – indeed overlapping -- linguistic theories for further discussion: 1) Biber’s multi-dimensional theory, 2) Ogura and Wang’s (1994) S-curve or ‘diffusion’ theory, 3) Kretzchmar’s (2009) linguistics of speech, and 4) Halliday’s (1987) notion of language as a dynamic open system. To summarise the outcomes of this thesis: 1) On variation between clausal types, it was shown that the distributional tendencies of verb types (sub, ind, mod) are different between the three adverbial clauses under consideration. 2) On variation between dialects, it has been shown that the northern area, i.e. the so-called Great Scandinavian Belt, displays an especially high comparative ratio of the inflexional subjunctive construction compared to the other areas. This thesis suggests that this result was caused by the influence of Norse, relating the finding to the argument of Samuels (1989) that the present tense -es ending in the northern dialect was introduced by the influence of the Scandinavians. 3) On variation between genres, those labelled Science, Documents and Religion display relatively high ratio of the inflexional subjunctive, while Letter, Romance and History show relatively low ratio of the inflexional subjunctive. This results are explained by Biber’s multi-dimensional theory, which shows that the inflexional subjunctive can be related to the factors ‘informational’, ‘non-narrative’, ‘persuasive’ and ‘abstract’. 4) Lastly, on the inflexional subjunctive in Late Middle English, this thesis concludes that 1) the change did not happen suddenly but gradually, and 2) the way language changes varies. Thus the inflexional subjunctive did not disappear suddenly from England, and there was a time lag among the clausal types, dialects and genres, which can be related to Ogura and Wang’s S-curve (“diffusion”) theory and Kretzchmars’s view of “linguistic continuum”. This thesis has shown that the issues with regard to the inflexional subjunctive are quite complex, so that research in this area requires not only textual analysis but also theoretical analysis, considering both intra- and extra- linguistic factors.