GYAN: A methodology for rule extraction from artificial neural networks

Artificial neural network (ANN) learning methods provide a robust and non-linear approach to approximating the target function for many classification, regression and clustering problems. ANNs have demonstrated good predictive performance in a wide variety of practical problems. However, there are strong arguments as to why ANNs are not sufficient for the general representation of knowledge. The arguments are the poor comprehensibility of the learned ANN, and the inability to represent explanation structures. The overall objective of this thesis is to address these issues by: (1) explanation of the decision process in ANNs in the form of symbolic rules (predicate rules with variables); and (2) provision of explanatory capability by mapping the general conceptual knowledge that is learned by the neural networks into a knowledge base to be used in a rule-based reasoning system. A multi-stage methodology GYAN is developed and evaluated for the task of extracting knowledge from the trained ANNs. The extracted knowledge is represented in the form of restricted first-order logic rules, and subsequently allows user interaction by interfacing with a knowledge based reasoner. The performance of GYAN is demonstrated using a number of real world and artificial data sets. The empirical results demonstrate that: (1) an equivalent symbolic interpretation is derived describing the overall behaviour of the ANN with high accuracy and fidelity, and (2) a concise explanation is given (in terms of rules, facts and predicates activated in a reasoning episode) as to why a particular instance is being classified into a certain category.

"Surviving" adolescence : apocalyptic and post-apocalyptic transformations in young adult fiction

This study, entitled "Surviving" Adolescence: Apocalyptic and post-apocalyptic transformations in young adult fiction‖, analyses how discourses surrounding the apocalyptic and post-apocalyptic are represented in selected young adult fiction published between 1997 and 2009. The term ―apocalypse‖ is used by current theorists to refer to an uncovering or disclosure (most often a truth), and ―post-apocalypse‖ means to be after a disclosure, after a revelation, or after catastrophe. This study offers a double reading of apocalyptic and post-apocalyptic discourses, and the dialectical tensions that are inherent in, and arise from, these discourses. Drawing on the current scholarship of children‘s and young adult literature this thesis uses post-structural theoretical perspectives to develop a framework and methodology for conducting a close textual analysis of exclusion, ‗un‘differentiation, prophecy, and simulacra of death. The combined theoretical perspectives and methodology offer new contributions to young adult fiction scholarship. This thesis finds that rather than conceiving adolescence as the endurance of a passing phase of a young person‘s life, there is a new trend emerging in young adult fiction that treats adolescence as a space of transformation essential to the survival of the young adult, and his/her community.

Generalised self-efficacy in relation to the life transitions of adult learners in a university setting : towards a narrative constructivist model of self-regulation

This paper demonstrates a model of self-regulation based on a qualitative research project with adult learners undertaking an undergraduate degree. The narrative about the participant’s life transitions, co-constructed with the researcher, yielded data about their generalised self-efficacy and resulted in a unique self-efficacy narrative for each participant. A model of self-regulation is proposed with potential applications for coaching, counselling and psychotherapy. A narrative method was employed to construct narratives about an individual’s self-efficacy in relation to their experience of learning and life transitions. The method involved a cyclical and iterative process using qualitative interviews to collect life history data from participants. In addition, research participants completed reflective homework tasks, and this data was included in the participant’s narratives. A highly collaborative method entailed narratives being co-constructed by researcher and research participants as the participants were guided in reflecting on their experience in relation to learning and life transitions; the reflection focused on behaviour, cognitions and emotions that constitute a sense of self-efficacy. The analytic process used was narrative analysis, in which life is viewed as constructed and experienced through the telling and retelling of stories and hence the analysis is the creation of a coherent and resonant story. The method of constructing self-efficacy narratives was applied to a sample of mature aged students starting an undergraduate degree. The research outcomes confirmed a three-factor model of self-efficacy, comprising three interrelated stages: initiating action, applying effort, and persistence in overcoming difficulties. Evaluation of the research process by participants suggested that they had gained an enhanced understanding of self-efficacy from their participation in the research process, and would be able to apply this understanding to their studies and other endeavours in the future. A model of self-regulation is proposed as a means for coaches, counsellors and psychotherapists working from a narrative constructivist perspective to assist clients facing life transitions by helping them generate selfefficacious cognitions, emotions and behaviour.

The application of neural networks to induction machine control

Neural networks (NNs) are discussed in connection with their possible use in induction machine drives. The mathematical model of the NN as well as a commonly used learning algorithm is presented. Possible applications of NNs to induction machine control are discussed. A simulation of an NN successfully identifying the nonlinear multivariable model of an induction-machine stator transfer function is presented. Previously published applications are discussed, and some possible future applications are proposed.

Identification and control of induction machines using artificial neural networks

The use of artificial neural networks (ANNs) to identify and control induction machines is proposed. Two systems are presented: a system to adaptively control the stator currents via identification of the electrical dynamics, and a system to adaptively control the rotor speed via identification of the mechanical and current-fed system dynamics. Both systems are inherently adaptive as well as self-commissioning. The current controller is a completely general nonlinear controller which can be used together with any drive algorithm. Various advantages of these control schemes over conventional schemes are cited, and the combined speed and current control scheme is compared with the standard vector control scheme

Identification and control of induction machines using artificial neural networks

This paper proposes the use of artificial neural networks (ANNs) to identify and control an induction machine. Two systems are presented: a system to adaptively control the stator currents via identification of the electrical dynamics; and a system to adaptively control the rotor speed via identification of the mechanical and current-fed system dynamics. Various advantages of these control schemes over other conventional schemes are cited and the performance of the combined speed and current control scheme is compared with that of the standard vector control scheme

Alleviating ‘overfitting’ via genetically-regularised neural network

A hybrid genetic algorithm/scaled conjugate gradient regularisation method is designed to alleviate ANN `over-fitting'. In application to day-ahead load forecasting, the proposed algorithm performs better than early-stopping and Bayesian regularisation, showing promising initial results.

Adult human articular chondrocytes in a microcarrier-based culture system : expansion and redifferentiation

xpanding human chondrocytes in vitro while maintaining their ability to form cartilage remains a key challenge in cartilage tissue engineering. One promising approach to address this is to use microcarriers as substrates for chondrocyte expansion. While microcarriers have shown beneficial effects for expansion of animal and ectopic human chondrocytes, their utility has not been determined for freshly isolated adult human articular chondrocytes. Thus, we investigated the proliferation and subsequent chondrogenic differentiation of these clinically relevant cells on porous gelatin microcarriers and compared them to those expanded using traditional monolayers. Chondrocytes attached to microcarriers within 2 days and remained viable over 4 weeks of culture in spinner flasks. Cells on microcarriers exhibited a spread morphology and initially proliferated faster than cells in monolayer culture, however, with prolonged expansion they were less proliferative. Cells expanded for 1 month and enzymatically released from microcarriers formed cartilaginous tissue in micromass pellet cultures, which was similar to tissue formed by monolayer-expanded cells. Cells left attached to microcarriers did not exhibit chondrogenic capacity. Culture conditions, such as microcarrier material, oxygen tension, and mechanical stimulation require further investigation to facilitate the efficient expansion of clinically relevant human articular chondrocytes that maintain chondrogenic potential for cartilage regeneration applications.

Enhancing pathways for ACE learners : Building the capacity of Adult and Community Education (ACE) providers to offer non-accredited courses and VET modules with pathways into VET programs.

This report presents findings from a project that considered a) the current capacity of Adult and Community Education (ACE) providers to offer non-accredited courses and single modules of accredited learning that provide pathways into full scale accredited VET programs, and b) the factors that aid and inhibit this from occurring. Based on the findings, suggestions are made as to what needs to be done to extend this capacity and thereby to achieve the goals outlined in the 2008 Ministerial Declaration on Adult Community Education.

Byron's back from the dead? : revenants of Romanticism in contemporary young adult fiction

Often identified as the origin of today’s children’s literature, Romanticism offers a particular context for interrogating boundaries between child and adult. Since the turn of the nineteenth century, however, Western society has “invented” the teenager to figure and to police the boundary between childhood and adulthood. In due course, twenty-first-century young adult (YA) novels such as Susan Davis’s Mad, Bad and Totally Dangerous (2004) and Cara Lockwood’s Wuthering High: A Bard Academy Novel (2006) have combined the Romantic and the adolescent in narratives which turn on supernatural invocation of Romantic authors as “really” present in contemporary adolescent lives. These novels tell stories of adolescence in which the self comes to be known via embodied encounters with dead authors, in particular, with Byron. Where “Byron scholarship has worked hard to disassociate the poet from this kind of pop-Gothic depiction, seeing it as the inevitable but regrettable offspring of nineteenth-century Byromania” (McDayter 30), contemporary YA fiction suggests that it is precisely via pop-Gothic depictions that today’s adolescents may first come to know the Romantic in general and the Byronic in particular. This paper reads these novels in the context of current anxieties about cultural illiteracy and educational “failure” in order to consider what work is being undertaken in the name of Byron, and to shed light on the ways in which cultural education may be taking place far beyond the realms of schools or cemeteries for today’s young readers.