AN AUTOPOIETIC APPROACH TO THE DEVELOPMENT OF SPEECH RECOGNITION (PENDEKATAN AUTOPOIETIC DALAM PEMBANGUNAN PENGECAMAN SUARA)

Abstract

The focus of research here is on the implementation of speech recognition through an autopoietic approach. The work done here has culminated in the introduction of a neural network architecture named Homunculus Network. This network was used in the development of a speech recognition system for Bahasa Melayu. The speech recognition system is an isolated-word, phoneme-level speech recognizer that is speaker independent and has a vocabulary of 15 words. The research done has identified some issues worth further work later. These issues are also the basis for the design and the development of the new autopoietic speech recognition system.