Over the past several years, I have been conducting research on subword modeling in speech recognition. The research is most specifically aimed at the difficult task of identifying and characterizing unknown words, although the proposed framework also has utility in other recognition tasks such as phonological and prosodic modeling. The approach exploits the linguistic substructure of words by ...

Amharic is the Semitic language that has the second large number of speakers after Arabic (Hayward and Richard 1999). Its writing system is syllabic with Consonant-Vowel (CV) syllable structure. Amharic orthography has more or less a one to one correspondence with syllabic sounds. We have used this feature of Amharic to develop a CV syllable-based speech recognizer, using Hidden Markov Modeling...

The research presented here examines the proposal that orthographic processing in reading polysyllabic words takes place via an analysis of the word into an orthographic/morphological structure called the Basic Orthographic Syllabic Structure or BOSS. This structure includes the largest possible coda in the first component (e.g., the THUND of THUNDER) and, as such, it cuts across the phonologic...

Utilizing corpora to build morphological analyzers for the purposes of computational application has been addressed in many different ways. Methods for automated morphological analysis generally focus on segmentation from raw text, and ignore the actual learning of what morpheme features are present. Other methods are time-consuming and require a great deal of prior knowledge of the language su...

Current-generation speech recognition systems seek to identify words via analysis of their underlying phonological constituents. Although this stratagem works well for carefully enunciated speech emanating from a pristine acoustic environment, it has fared less well for recognizing speech spoken under more realistic conditions, such as (1) moderate to high levels of background noise (2) moderat...

This contribution explores the extent to which rhythm-based features of poetic texts can contribute meaningfully authorship recognition. We show that, although a binary categorization languages as syllabic vs. accentual-syllabic fails fully explain differences. However, once we formalize accentual regularity continuum, our analysis shows that attribution results improve move from most least acc...

In 1968, Ernst Pulgram began his classic monograph on the syllable with the wise words, “conscience, courtesy, and caution require that anyone wishing to concern himself with the syllable read all, or at least most, of the enormous literature on it.” The years since his study have only magnified the challenge of this suggestion. Yet life is short, and space shorter still, and so in these few pa...

The auditory cortex communicates with the frontal lobe via the middle temporal gyrus (auditory ventral stream; AVS) or the inferior parietal lobule (auditory dorsal stream; ADS). Whereas the AVS is ascribed only with sound recognition, the ADS is ascribed with sound localization, voice detection, prosodic perception/production, lip-speech integration, phoneme discrimination, articulation, repet...

This paper investigates the potential role of syllabic structure in characterising the informational content of running speech using an energy-based measure (the cochlea-scaled entropy, CSE index). We computed the CSE and compared how it aligned to the energy envelope for a corpus of English and Spanish sentences. We also compared these measures to syllabic structure, which differs markedly bet...