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.

The 'moving targets' training algorithm

A simple method for training the dynamical behavior of a neural network is derived. It is applicable to any training problem in discrete-time networks with arbitrary feedback. The algorithm resembles back-propagation in that an error function is minimized using a gradient-based method, but the optimization is carried out in the hidden part of state space either instead of, or in addition to weight space. Computational results are presented for some simple dynamical training problems, one of which requires response to a signal 100 time steps in the past.

The 'moving targets' training algorithm

A simple method for training the dynamical behavior of a neural network is derived. It is applicable to any training problem in discrete-time networks with arbitrary feedback. The method resembles back-propagation in that it is a least-squares, gradient-based optimization method, but the optimization is carried out in the hidden part of state space instead of weight space. A straightforward adaptation of this method to feedforward networks offers an alternative to training by conventional back-propagation. Computational results are presented for simple dynamical training problems, with varied success. The failures appear to arise when the method converges to a chaotic attractor. A patch-up for this problem is proposed. The patch-up involves a technique for implementing inequality constraints which may be of interest in its own right.

Information geometric measurements of generalisation

Neural networks can be regarded as statistical models, and can be analysed in a Bayesian framework. Generalisation is measured by the performance on independent test data drawn from the same distribution as the training data. Such performance can be quantified by the posterior average of the information divergence between the true and the model distributions. Averaging over the Bayesian posterior guarantees internal coherence; Using information divergence guarantees invariance with respect to representation. The theory generalises the least mean squares theory for linear Gaussian models to general problems of statistical estimation. The main results are: (1)~the ideal optimal estimate is always given by average over the posterior; (2)~the optimal estimate within a computational model is given by the projection of the ideal estimate to the model. This incidentally shows some currently popular methods dealing with hyperpriors are in general unnecessary and misleading. The extension of information divergence to positive normalisable measures reveals a remarkable relation between the dlt dual affine geometry of statistical manifolds and the geometry of the dual pair of Banach spaces Ld and Ldd. It therefore offers conceptual simplification to information geometry. The general conclusion on the issue of evaluating neural network learning rules and other statistical inference methods is that such evaluations are only meaningful under three assumptions: The prior P(p), describing the environment of all the problems; the divergence Dd, specifying the requirement of the task; and the model Q, specifying available computing resources.

Incorporating curvature information into on-line learning

We analyse the dynamics of a number of second order on-line learning algorithms training multi-layer neural networks, using the methods of statistical mechanics. We first consider on-line Newton's method, which is known to provide optimal asymptotic performance. We determine the asymptotic generalization error decay for a soft committee machine, which is shown to compare favourably with the result for standard gradient descent. Matrix momentum provides a practical approximation to this method by allowing an efficient inversion of the Hessian. We consider an idealized matrix momentum algorithm which requires access to the Hessian and find close correspondence with the dynamics of on-line Newton's method. In practice, the Hessian will not be known on-line and we therefore consider matrix momentum using a single example approximation to the Hessian. In this case good asymptotic performance may still be achieved, but the algorithm is now sensitive to parameter choice because of noise in the Hessian estimate. On-line Newton's method is not appropriate during the transient learning phase, since a suboptimal unstable fixed point of the gradient descent dynamics becomes stable for this algorithm. A principled alternative is to use Amari's natural gradient learning algorithm and we show how this method provides a significant reduction in learning time when compared to gradient descent, while retaining the asymptotic performance of on-line Newton's method.

Mixture density network training by computation in parameter space

Training Mixture Density Network (MDN) configurations within the NETLAB framework takes time due to the nature of the computation of the error function and the gradient of the error function. By optimising the computation of these functions, so that gradient information is computed in parameter space, training time is decreased by at least a factor of sixty for the example given. Decreased training time increases the spectrum of problems to which MDNs can be practically applied making the MDN framework an attractive method to the applied problem solver.

Delivering phonological and phonics training within whole-class teaching

Background: Early, intensive phonological awareness and phonics training is widely held to be beneficial for children with poor phonological awareness. However, most studies have delivered this training separately from children's normal whole-class reading lessons. Aims: We examined whether integrating this training into whole class, mixed-ability reading lessons could impact on children with poor phonological awareness, whilst also benefiting normally developing readers. Sample: Teachers delivered the training within a broad reading programme to whole classes of children from Reception to the end of Year 1 (N=251). A comparison group of children received standard teaching methods (N=213). Method: Children's literacy was assessed at the beginning of Reception, and then at the end of each year until 1 year post-intervention. Results: The strategy significantly impacted on reading performance for normally developing readers and those with poor phonological awareness, vastly reducing the incidence of reading difficulties from 20% in comparison schools to 5% in intervention schools. Conclusions: Phonological and phonics training is highly effective for children with poor phonological awareness, even when incorporated into whole-class teaching.

ALEX:mobile language assistant for low-literacy adults

Basic literacy skills are fundamental building blocks of education, yet for a very large number of adults tasks such as understanding and using everyday items is a challenge. While research, industry, and policy-making is looking at improving access to textual information for low-literacy adults, the literacy-based demands of today's society are continually increasing. Although many community-based organizations offer resources and support to adults with limited literacy skills, current programs have difficulties reaching and retaining those that would benefit most from them. To address these challenges, the National Research Council of Canada is proposing a technological solution to support literacy programs and to assist low-literacy adults in today's information-centric society: ALEX© – Adult Literacy support application for EXperiential learning. ALEX© has been created together with low-literacy adults, following guidelines for inclusive design of mobile assistive tools. It is a mobile language assistant that is designed to be used both in the classroom and in daily life, in order to help low-literacy adults become increasingly literate and independent.

Self-training from labeled features for sentiment analysis

Sentiment analysis concerns about automatically identifying sentiment or opinion expressed in a given piece of text. Most prior work either use prior lexical knowledge defined as sentiment polarity of words or view the task as a text classification problem and rely on labeled corpora to train a sentiment classifier. While lexicon-based approaches do not adapt well to different domains, corpus-based approaches require expensive manual annotation effort. In this paper, we propose a novel framework where an initial classifier is learned by incorporating prior information extracted from an existing sentiment lexicon with preferences on expectations of sentiment labels of those lexicon words being expressed using generalized expectation criteria. Documents classified with high confidence are then used as pseudo-labeled examples for automatical domain-specific feature acquisition. The word-class distributions of such self-learned features are estimated from the pseudo-labeled examples and are used to train another classifier by constraining the model's predictions on unlabeled instances. Experiments on both the movie-review data and the multi-domain sentiment dataset show that our approach attains comparable or better performance than existing weakly-supervised sentiment classification methods despite using no labeled documents.

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.

A novel framework of training hidden Markov support vector machines from lightly-annotated data

Natural language understanding (NLU) aims to map sentences to their semantic mean representations. Statistical approaches to NLU normally require fully-annotated training data where each sentence is paired with its word-level semantic annotations. In this paper, we propose a novel learning framework which trains the Hidden Markov Support Vector Machines (HM-SVMs) without the use of expensive fully-annotated data. In particular, our learning approach takes as input a training set of sentences labeled with abstract semantic annotations encoding underlying embedded structural relations and automatically induces derivation rules that map sentences to their semantic meaning representations. The proposed approach has been tested on the DARPA Communicator Data and achieved 93.18% in F-measure, which outperforms the previously proposed approaches of training the hidden vector state model or conditional random fields from unaligned data, with a relative error reduction rate of 43.3% and 10.6% being achieved.

Standardisation and benchmarking for improving translator training

The worldwide growth of the translation industry requires qualified professional translators. During the last decade, we have seen an enormous increase in translator training programmes offered by universities, mainly at postgraduate level. A challenge for such university programmes is to make sure that they prepare graduates who are qualified for the needs of the diverse profession in the rapidly changing market. This means that programmes need to be developed with the market needs in mind and that they need to ensure a good match between graduates' competences and employers' requirements. This paper addresses the following questions: How can universities adapt translator training programmes to the rapidly changing industry and the accompanying changes in professional profiles? How can we reconcile the requests of the industry for graduates who have practical and professional skills with the requests of the universities for graduates who have in-depth academic knowledge and intellectual skills? What standards and benchmarks are in place to assure quality of translator training programmes? Some such developments in respect of benchmarking are illustrated at first for the United Kingdom, followed by information on the European Master’s in Translation (EMT) project, an initiative at the European level. Finally, the paper reflects on the challenges which the EMT translator competence profile poses for university programmes.

Hidden in plain sight:low-literacy adults in a developed country overcoming social and educational challenges through mobile learning support tools

Illiteracy is often associated with people in developing countries. However, an estimated 50 % of adults in a developed country such as Canada lack the literacy skills required to cope with the challenges of today's society; for them, tasks such as reading, understanding, basic arithmetic, and using everyday items are a challenge. Many community-based organizations offer resources and support for these adults, yet overall functional literacy rates are not improving. This is due to a wide range of factors, such as poor retention of adult learners in literacy programs, obstacles in transferring the acquired skills from the classroom to the real life, personal attitudes toward learning, and the stigma of functional illiteracy. In our research we examined the opportunities afforded by personal mobile devices in providing learning and functional support to low-literacy adults. We present the findings of an exploratory study aimed at investigating the reception and adoption of a technological solution for adult learners. ALEX© is a mobile application designed for use both in the classroom and in daily life in order to help low-literacy adults become increasingly literate and independent. Such a solution complements literacy programs by increasing users' motivation and interest in learning, and raising their confidence levels both in their education pursuits and in facing the challenges of their daily lives. We also reflect on the challenges we faced in designing and conducting our research with two user groups (adults enrolled in literacy classes and in an essential skills program) and contrast the educational impact and attitudes toward such technology between these. Our conclusions present the lessons learned from our evaluations and the impact of the studies' specific challenges on the outcome and uptake of such mobile assistive technologies in providing practical support to low-literacy adults in conjunction with literacy and essential skills training. © 2013 Her Majesty the Queen in Right of Canada.

Training staff in voluntary organisations:challenges in providing a service to adults with learning disabilities

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

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