Context-free and context-sensitive dynamics in recurrent neural networks

Continuous-valued recurrent neural networks can learn mechanisms for processing context-free languages. The dynamics of such networks is usually based on damped oscillation around fixed points in state space and requires that the dynamical components are arranged in certain ways. It is shown that qualitatively similar dynamics with similar constraints hold for a(n)b(n)c(n), a context-sensitive language. The additional difficulty with a(n)b(n)c(n), compared with the context-free language a(n)b(n), consists of 'counting up' and 'counting down' letters simultaneously. The network solution is to oscillate in two principal dimensions, one for counting up and one for counting down. This study focuses on the dynamics employed by the sequential cascaded network, in contrast to the simple recurrent network, and the use of backpropagation through time. Found solutions generalize well beyond training data, however, learning is not reliable. The contribution of this study lies in demonstrating how the dynamics in recurrent neural networks that process context-free languages can also be employed in processing some context-sensitive languages (traditionally thought of as requiring additional computation resources). This continuity of mechanism between language classes contributes to our understanding of neural networks in modelling language learning and processing.

An object-oriented neural network approach to short-term traffic forecasting

This paper discusses an object-oriented neural network model that was developed for predicting short-term traffic conditions on a section of the Pacific Highway between Brisbane and the Gold Coast in Queensland, Australia. The feasibility of this approach is demonstrated through a time-lag recurrent network (TLRN) which was developed for predicting speed data up to 15 minutes into the future. The results obtained indicate that the TLRN is capable of predicting speed up to 5 minutes into the future with a high degree of accuracy (90-94%). Similar models, which were developed for predicting freeway travel times on the same facility, were successful in predicting travel times up to 15 minutes into the future with a similar degree of accuracy (93-95%). These results represent substantial improvements on conventional model performance and clearly demonstrate the feasibility of using the object-oriented approach for short-term traffic prediction. (C) 2001 Elsevier Science B.V. All rights reserved.

Small area population disease burden

Small area health statistics has assumed increasing importance as the focus of population and public health moves to a more individualised approach of smaller area populations. Small populations and low event occurrence produce difficulties in interpretation and require appropriate statistical methods, including for age adjustment. There are also statistical questions related to multiple comparisons. Privacy and confidentiality issues include the possibility of revealing information on individuals or health care providers by fine cross-tabulations. Interpretation of small area population differences in health status requires consideration of migrant and Indigenous composition, socio-economic status and rural-urban geography before assessment of the effects of physical environmental exposure and services and interventions. Burden of disease studies produce a single measure for morbidity and mortality - disability adjusted life year (DALY) - which is the sum of the years of life lost (YLL) from premature mortality and the years lived with disability (YLD) for particular diseases (or all conditions). Calculation of YLD requires estimates of disease incidence (and complications) and duration, and weighting by severity. These procedures often mean problematic assumptions, as does future discounting and age weighting of both YLL and YLD. Evaluation of the Victorian small area population disease burden study presents important cross-disciplinary challenges as it relies heavily on synthetic approaches of demography and economics rather than on the empirical methods of epidemiology. Both empirical and synthetic methods are used to compute small area mortality and morbidity, disease burden, and then attribution to risk factors. Readers need to examine the methodology and assumptions carefully before accepting the results.

A preliminary phylogeny of the scale insects (Hemiptera : Sternorrhyncha : Coccoidea) based on nuclear small-subunit ribosomal DNA

Scale insects (Hemiptera: Sternorrhyncha: Coccoidea) are a speciose and morphologically specialized group of plant-feeding bugs in which evolutionary relationships and thus higher classification are controversial. Sequences derived from nuclear small-subunit ribosomal DNA were used to generate a preliminary molecular phylogeny for the Coccoidea based on 39 species representing 14 putative families. Monophyly of the archaeococcoids (comprising Ortheziidae, Margarodidae sensu lato, and Phenacoleachia) was equivocal, whereas monophyly of the neococcoids was supported. Putoidae, represented by Puto yuccae, was found to be outside the remainder of the neococcoid clade. These data are consistent with a single origin (in the ancestor of the neococcoid clade) of a chromosome system involving paternal genome elimination in males. Pseudococcidae (mealybugs) appear to be sister to the rest of the neococcoids and there are indications that Coccidae (soft scales) and Kerriidae (lac scales) are sister taxa. The Eriococcidae (felt scales) was not recovered as a monophyletic group and the eriococcid genus Eriococcus sensu lato was polyphyletic. (C) 2002 Elsevier Science (USA). All rights reserved.

Inter-subband redundancy prediction using neural network for video coding

High performance video codec is mandatory for multimedia applications such as video-on-demand and video conferencing. Recent research has proposed numerous video coding techniques to meet the requirement in bandwidth, delay, loss and Quality-of-Service (QoS). In this paper, we present our investigations on inter-subband self-similarity within the wavelet-decomposed video frames using neural networks, and study the performance of applying the spatial network model to all video frames over time. The goal of our proposed method is to restore the highest perceptual quality for video transmitted over a highly congested network. Our contributions in this paper are: (1) A new coding model with neural network based, inter-subband redundancy (ISR) prediction for video coding using wavelet (2) The performance of 1D and 2D ISR prediction, including multiple levels of wavelet decompositions. Our result shows a short-term quality enhancement may be obtained using both 1D and 2D ISR prediction.