Description and training of neural network dynamics

Attractor properties of a popular discrete-time neural network model are illustrated through numerical simulations. The most complex dynamics is found to occur within particular ranges of parameters controlling the symmetry and magnitude of the weight matrix. A small network model is observed to produce fixed points, limit cycles, mode-locking, the Ruelle-Takens route to chaos, and the period-doubling route to chaos. Training algorithms for tuning this dynamical behaviour are discussed. Training can be an easy or difficult task, depending whether the problem requires the use of temporal information distributed over long time intervals. Such problems require training algorithms which can handle hidden nodes. The most prominent of these algorithms, back propagation through time, solves the temporal credit assignment problem in a way which can work only if the relevant information is distributed locally in time. The Moving Targets algorithm works for the more general case, but is computationally intensive, and prone to local minima.

Efficient training of RBF networks for classification.

Radial Basis Function networks with linear outputs are often used in regression problems because they can be substantially faster to train than Multi-layer Perceptrons. For classification problems, the use of linear outputs is less appropriate as the outputs are not guaranteed to represent probabilities. We show how RBFs with logistic and softmax outputs can be trained efficiently using the Fisher scoring algorithm. This approach can be used with any model which consists of a generalised linear output function applied to a model which is linear in its parameters. We compare this approach with standard non-linear optimisation algorithms on a number of datasets.

Bayesian training of mixture density networks

Mixture Density Networks (MDNs) are a well-established method for modelling the conditional probability density which is useful for complex multi-valued functions where regression methods (such as MLPs) fail. In this paper we extend earlier research of a regularisation method for a special case of MDNs to the general case using evidence based regularisation and we show how the Hessian of the MDN error function can be evaluated using R-propagation. The method is tested on two data sets and compared with early stopping.

Efficient training of RBF networks for classification

Radial Basis Function networks with linear outputs are often used in regression problems because they can be substantially faster to train than Multi-layer Perceptrons. For classification problems, the use of linear outputs is less appropriate as the outputs are not guaranteed to represent probabilities. In this paper we show how RBFs with logistic and softmax outputs can be trained efficiently using algorithms derived from Generalised Linear Models. This approach is compared with standard non-linear optimisation algorithms on a number of datasets.

Teachers' conceptions of their role

An analysis is made of the conceptions which serving teachers have of their role, though no attempt is made to relate this to their practice of teaching. A series of role items was collected to afford a description of the teacher's role in terms of school and society expectations as well as classroom behaviours. These were taken from the literature and from interviews with teachers, and confirmed in a preliminary survey. Presented as a questionnaire, replies to the main investigation were made by 881 teachers, working in a variety of schools from nurseries to comprehensives. Two attempts have been made to construct a role model. The first, depending on the judgement of items fitting theoretically derived roles, failed, due to diffuseness in the role of teacher. The second used factor analysis; six factors were extracted which represent meaningful and distinct areas of role. The analysis has depended largely on examination of scores taken from these factors. Teachers in all types of school have similar conceptions of discipline. Nursery-infant and junior staff generally agree on the other areas investigated, but the concepts of secondary teachers are distinct. They are more conservative and less child-centered. When the class being taught is held constant, few differences in role conception are found to be related to sex, being a parent, graduate status, or personality, as measured in terms of the extrovert and neurotic dimensions. The first few years of teaching bring considerable changes in role conception, and further changes occur with prolonged experience. Deputy heads in junior schools and nursery nurses have quite distinct role conceptions; those of all other teachers, including those holding senior posts in secondary schools, are similar. The perception of school climate influences the role conception of primary teachers directly, but it does not influence that of secondary teachers. The greatest variation in role conception is related to scores on the radical scale of Oliver and Butcher. Primary school teachers experience little constraint, but that reported by secondary school teachers is considerable, especially that coming from the head. Despite difficulties caused by the wide division between primary and secondary education, teachers have an accurate perception of the roles their colleagues adopt. A few misunderstandings may be due to a feeling of idealism amongst nursery and infant teachers. There is evidence in their conception of role that would enhance the professional standing of teachers, but this is not in a form which is likely to be recognised by the public.

Discriminative training of the hidden vector state model for semantic parsing

In this paper, we discuss how discriminative training can be applied to the hidden vector state (HVS) model in different task domains. The HVS model is a discrete hidden Markov model (HMM) in which each HMM state represents the state of a push-down automaton with a finite stack size. In previous applications, maximum-likelihood estimation (MLE) is used to derive the parameters of the HVS model. However, MLE makes a number of assumptions and unfortunately some of these assumptions do not hold. Discriminative training, without making such assumptions, can improve the performance of the HVS model by discriminating the correct hypothesis from the competing hypotheses. Experiments have been conducted in two domains: the travel domain for the semantic parsing task using the DARPA Communicator data and the Air Travel Information Services (ATIS) data and the bioinformatics domain for the information extraction task using the GENIA corpus. The results demonstrate modest improvements of the performance of the HVS model using discriminative training. In the travel domain, discriminative training of the HVS model gives a relative error reduction rate of 31 percent in F-measure when compared with MLE on the DARPA Communicator data and 9 percent on the ATIS data. In the bioinformatics domain, a relative error reduction rate of 4 percent in F-measure is achieved on the GENIA corpus.

The use of open learning in the training of neurotechnologists:a regional case study

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

The education and training of technicians in the construction industry

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

The effectiveness of an organisations' equal opportunities policy for the role and training of women, particularly in the fire service

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