It has become widely accepted that temporal asymmetry is largely a cosmological problem; the task of explaining temporal asymmetry reduces in the main to that of explaining an aspect of the condition of the early universe. However, cosmologists who discuss these issues often make mistakes similar to those that plagued nineteenth century discussions of the statistical foundations of thermodynami...

Good quality base and sub-base material for road pavement construction is becoming increasingly scarce. Dolerite is one of the more abundant natural road building materials in South Africa. Yet, it seems that the road design and construction fraternity shy away from using it in favour of more expensive commercially produced crushed material – and so artificially increase the scarcity of natural...

The subword complexity of an infinite word ξ is a function f (ξ, n) returning the number of finite subwords (factors, infixes) of length n of ξ. In the present paper we investigate infinite words for which the set of subwords occurring infinitely often is a regular language. Among these infinite words we characterise those which are eventually recurrent. Furthermore, we derive some results comp...

For speech recognition with large vocabularies, a user should not be burdened with having to train several thousand words explicitly. Therefore, it proves extremely useful to provide a means for easy vocabulary generation and enlargement from written text input. Applying a set of appropriately defined rules, the orthography of a lexicon item is first transcribed into the phonetic symbols of the...

Neural machine translation together with subword segmentation has recently produced state-of-the-art translation performance. The commonly used segmentation algorithm based on byte-pair encoding (BPE) does not consider the morphological structure of words. This occasionally causes misleading segmentation and incorrect translation of rare words. In this thesis we explore the use of morphological...

A major challenge in speech recognition is creating a lexicon which is robust to inter-and intra-speaker variations. This is even more so in speech recognisers based on non-linguistic units, e.g., acoustic subword units (ASWUs), since no standard pronunciation dictionaries are available. Thus the baseforms describing the vocabulary words in terms of the recognition units need to be generated fr...

An architecture for achieving word recognition and incremental learning of new words in a language processing system is presented. The architecture is based on neural associative memories and hidden Markov models. The hidden Markov models generate subword-unit transcriptions of the spoken words and provide them as input to the associative memory module. The associative memory module is a networ...

isfaction neural network (CSNN) model developed for In this paper, we address the issues in recognition of a large number of subword units of speech with high confusability among several units. Evidence available from the classification models trained with a limited number of training examples may not be strong to correctly recognize the subword units. We present a constraint satisfaction neura...

Reversing arrows in the classic tri-variate X-M-Y mediation models as a test to check whether one mediation model is superior to another is inadmissible. Presenting evidence that one trivariate mediation model yields a significant indirect effect, whereas one with some reversed arrows does not, is not proof or even evidence that one model should be preferred. In fact, the significance of the in...