The n-tuple network: coping with sub-optimality

The subject of this thesis is the n-tuple net.work (RAMnet). The major advantage of RAMnets is their speed and the simplicity with which they can be implemented in parallel hardware. On the other hand, this method is not a universal approximator and the training procedure does not involve the minimisation of a cost function. Hence RAMnets are potentially sub-optimal. It is important to understand the source of this sub-optimality and to develop the analytical tools that allow us to quantify the generalisation cost of using this model for any given data. We view RAMnets as classifiers and function approximators and try to determine how critical their lack of' universality and optimality is. In order to understand better the inherent. restrictions of the model, we review RAMnets showing their relationship to a number of well established general models such as: Associative Memories, Kamerva's Sparse Distributed Memory, Radial Basis Functions, General Regression Networks and Bayesian Classifiers. We then benchmark binary RAMnet. model against 23 other algorithms using real-world data from the StatLog Project. This large scale experimental study indicates that RAMnets are often capable of delivering results which are competitive with those obtained by more sophisticated, computationally expensive rnodels. The Frequency Weighted version is also benchmarked and shown to perform worse than the binary RAMnet for large values of the tuple size n. We demonstrate that the main issues in the Frequency Weighted RAMnets is adequate probability estimation and propose Good-Turing estimates in place of the more commonly used :Maximum Likelihood estimates. Having established the viability of the method numerically, we focus on providillg an analytical framework that allows us to quantify the generalisation cost of RAMnets for a given datasetL. For the classification network we provide a semi-quantitative argument which is based on the notion of Tuple distance. It gives a good indication of whether the network will fail for the given data. A rigorous Bayesian framework with Gaussian process prior assumptions is given for the regression n-tuple net. We show how to calculate the generalisation cost of this net and verify the results numerically for one dimensional noisy interpolation problems. We conclude that the n-tuple method of classification based on memorisation of random features can be a powerful alternative to slower cost driven models. The speed of the method is at the expense of its optimality. RAMnets will fail for certain datasets but the cases when they do so are relatively easy to determine with the analytical tools we provide.

Self-attentional processes in anxiety:an experimental study

Self-attention research has demonstrated a relationship between dispositional self-focus, anxiety proneness and fear arousal. In addition, the effect of self-focus manipulations on approach-avoidance tasks involving a feared stimulus are strikingly similar to the effects obtained from manipulation of other cognitive factors such as perceived self-efficacy. A number of experiments were designed to explore the relationship between self-focused attention and ffilxiety. Data from the experiments demonstrate that self-attention influences a variety of cognitive variables which have been considered as central factors in anxiety. Concomitants of self-focus are increased awareness of physiological arousal and overestimation of such arousal, the identification of self-discrepancies, cognitive failures and performance deficits and the activation of physical threat concepts in memory. These factors are conceptualised as central in the negative evaluation of physiological arousal and coping resources in anxiety. Clinically anxious individuals typically have high scores in dispositional self-consciousness and body-consciousness. In patients suffering from generalised anxiety or panic disorders maladaptive self-focusing tendencies can be related to specific life stressors which render aspects of the self salient. An analysis of the ideational component of anxiety revealed three subcomponents; negative social ideation (worry about other people's reaction to the self), negative somatic ideation (worry about physical symptoms and health) and obsessional ideation (the experience of uncontrollable and repetitive thoughts) which were differentially associated with measures of dispositional self-focus. The frequency and content of an.xious w-orry is associated with specific self-focusing tendencies. It is proposed that the 'attentional style' of the individual is an important determinant of the nature and intensity of their affective response in a threatening situation. A self-attentional model of anxiety is proposed and the complex interaction between self-focus and other cognitive factors in anxiety such as appraisal of arousal and coping resources and perceived levels of self-efficacy is discussed. The model presents new directions for research and therapeutic intervention in anxiety.

Perception of risk, feelings of vulnerability and effects on driving behaviour in women drivers across the lifespan.

Sustained driving in older age has implications for quality of life and mental health. Studies have shown that despite the recognised importance of driving in maintaining health and social engagement, many women give up driving prematurely or adopt self-imposed restrictive driving practices. Emotional responses to driving have been implicated in these decisions. This research examined the effect of risk perception and feelings of vulnerability on women’s driving behaviour across the lifespan. It also developed and tested a modified theory of planned behaviour intervention to positively affect driving habits. The first two studies (N=395) used quantitative analysis to model driving behaviours affected by risk perception and feelings of vulnerability, and established that feelings of vulnerability do indeed affect women’s driving behaviour, specifically resulting in increases in driving avoidance and the adoption of maladaptive driving styles. Further, that self-regulation, conceptualised as avoidance, is used by drivers across the lifespan. Qualitative analysis of focus group data (N=48) in the third study provided a deeper understanding of the variations in coping behaviours adopted by sub-groups of drivers and extended the definition of self-regulation to incorporate adaptive coping strategies. The next study (N=64) reported the construction and preliminary validation of the novel self-regulation index (SRI) to measure wider self-regulation behaviours using an objective measure of driving behaviour, a simulated driving task. The understanding gained from the formative research was used in the final study, an extended theory of planned behaviour intervention to promote wider self-regulation behaviour, measured using the previously validated self-regulation index. The intervention achieved moderate success with changes in affective attitude and normative beliefs as well as self-reported behaviour. The results offer promise for self-regulation, incorporating a spectrum of planning and coping behaviours, to be used as a mechanism to assist drivers in achieving their personal mobility goals whilst promoting safe driving.

Attributions, cognitions, and coping styles:teleworkers' reactions to work-related problems

Based on the attributional reformulation of learned helplessness theory (Abramson, Seligman, & Teasdale, 1978) and Lazarus and Launier's (1978) primary-secondary appraisal theory of stress, the present study sought to examine teleworkers' reactions to their work-related problems. The role of attributions about the sources, and cognitions about the consesquences, of these problems in promoting positive adaptation was addressed. In particular, it was predicted that teleworkers who made optimistic attributions and cognitions would be more likely to employ problem-focused coping strategies and, as a result, report more positive psychological and job-related outcomes. Based on a survey sample of 192 teleworkers, the results indicated that a tendency to engage in self-blame was related to the use of emotion-focused coping strategies. In turn, there was evidence linking emotion-focused coping strategies to negative outcomes and problem-focused coping strategies to positive outcomes. The results are discussed in relation to attributional approaches to stress which highlight the importance of cognitions about the consequences of negative events. Finally, implications for the training of teleworkers are presented.

Relaxing claims:coping with uncertainty while evaluating assumptions at run time

Self-adaptation enables software systems to respond to changing environmental contexts that may not be fully understood at design time. Designing a dynamically adaptive system (DAS) to cope with this uncertainty is challenging, as it is impractical during requirements analysis and design time to anticipate every environmental condition that the DAS may encounter. Previously, the RELAX language was proposed to make requirements more tolerant to environmental uncertainty, and Claims were applied as markers of uncertainty that document how design assumptions affect goals. This paper integrates these two techniques in order to assess the validity of Claims at run time while tolerating minor and unanticipated environmental conditions that can trigger adaptations. We apply the proposed approach to the dynamic reconfiguration of a remote data mirroring network that must diffuse data while minimizing costs and exposure to data loss. Results show RELAXing Claims enables a DAS to reduce adaptation costs. © 2012 Springer-Verlag.