Narrowing the gap between automatic and human word recognition

In everyday life, speech is all around us, on the radio, television, and in human-human interaction. We are continually confronted with novel utterances, and usually we have no problem recognising and understanding them. Several research fields investigate the speech recognition process. This thesis addresses two of these: human speech recognition (HSR) and automatic speech recognition (ASR). Although these research fields appear closely related, their aims and research methods are quite different. Nevertheless, despite the gap that has separated ASR and HSR for decades, there is a growing interest in possible cross-fertilisation. The central goal of this thesis is to narrow the gap between ASR and HSR. The focus of this thesis is on word recognition given its central role in both fields. The research described in this thesis demonstrates the close parallels between ASR and HSR at the level of the computational processes involved in word recognition. These parallels are illustrated by the development of the end-to-end computational model of HSR, created with techniques from the field of ASR. This model is called SpeM. This thesis will interest researchers in the fields of human and automatic speech recognition, word recognition, computational modelling, and phonetic