Standard hidden Markov model (HMM) based automatic speech recognition (ASR) systems usually use cepstral features as acoustic observation and phonemes as subword units. Speech signal exhibits wide range of variability such as, due to environmental variation, speaker variation. This leads to different kinds of mismatch, such as, mismatch between acoustic features and acoustic models or mismatch ...

The Neural Prediction Model is the speech recognition model based on pattern prediction by multilayer perceptrons. Its effectiveness was confirmed by the speaker-independent digit recognition experiments. This paper presents an improvement in the model and its application to large vocabulary speech recognition, based on subword units. The improvement involves an introduction of "backward predic...

This paper describes the development and application of a phonetic recognition system to the task of spoken document retrieval. The recognizer is used to generate phonetic transcriptions of the speech messages which are then processed to produce subword unit representations for indexing and retrieval. Subword units are used as an alternative to words units generated by either keyword spotting o...

We show that microSIMD architectures are more efficient for media processing than other parallel architectures like SIMD or MIMD parallel processor architectures, and VLIW or superscalar architectures. We define alternative mappings of data onto subwords, and show that the index mapping is an ideal mapping for achieving maximal subword parallelism with minimal revamping of the original serial l...

The joint spectral radius of a bounded set of d × d real matrices is defined to be the maximum possible exponential growth rate of products of matrices drawn from that set. For a fixed set of matrices, a sequence of matrices drawn from that set is called extremal if the associated sequence of partial products achieves this maximal rate of growth. An influential conjecture of J. Lagarias and Y. ...

In this paper we explore various interconnections between rich words, Sturmian words, and trapezoidal words. Rich words, first introduced by the second and third authors together with J. Justin and S. Widmer, constitute a new class of finite and infinite words characterized by having the maximal number of palindromic factors. Every finite Sturmian word is rich, but not conversely. Trapezoidal w...

This paper is an overview of results' on the subword complexity of various classes of languages obtained by iterating a mapping, and more precisely on the asymptotic behaviour of this complexity. In the first section we consider iterated morphisms, that is DOL languages. In this area the subword complexity falls into one of five classes, and class membership can be determined. The second sectio...