Biologically-Motivated Machine Learning of Natural Language and Ontology - A Computational Cognitive Model

Abstract:

The individual cognitive science disciplines all have contributions to make to the understanding and modelling of human learning. Our previous research has explored unsupervised learning of phonology, morphology and low-level syntax, as well as basic noun, verb and preposition ontology and semantics, plus musical and speech prosody. Successful applications using a mix of supervised and unsupervised techniques include speech control of equipment, deep web search, confused word spelling correction, multi-lingual semantic models and audio-visual speech recognition.

Our current research is focused on doing simultaneous learning of ontology, syntax and semantics by embedding the learner in realistic situations and by developing low-level biologically-plausible models of perceptual and cognitive processing.

Keywords: Ontology, Ontogenesis, Stereology, Stereophony, Stereopsis, Stereognosis, Stereosemy, Cognitive Linguistics, Computational Linguistics, Sensor Fusion, Data Fusion, Neurological Models, Neural Nets.

Parking: Visitors requiring a parking pass are asked to contact us at least one working day before the seminar.

Enquiries: sals@ics.mq.edu.au