Events Calendar

Seminar -

Speaker: Odette Scharenborg, Radboud University (Nijmegen, the Netherlands)

Talk Title: Modeling and Understanding Human Spoken-Word Recognition

Abstract: The question that underlies most of my research is the question why humans are so much better at recognizing speech than computers. I have approached this question from several angles, from the field of automatic speech recognition, the field of psycholinguistics, and through the combination of the two, i.e., the computational modeling of human spoken-word recognition. In this talk, I will present results from my computational modelling and psycholinguistics work.

In the first part, I will present my computational model, which is able to recognize real speech, Fine-Tracker. Fine-Tracker was specifically developed to account for the accumulating evidence that subtle phonetic detail in the speech signal is important in human spoken-word recognition. I will explain the model and illustrate its modelling ability by presenting a simulation study investigating the role of durational information in resolving temporary ambiguity due to lexical embedding (i.e., 'ham' in the longer word 'hamster') to aid spoken-word recognition. I will start the talk by briefly discussing the value of computational modelling in spoken-word recognition.

In the second part of this talk, I will focus on the results obtained in my current project on human non-native word recognition in noise. Most people will have noticed that communication in the presence of background noise is more difficult in a non-native than in the native language - even for those who have a high proficiency in the non-native language involved. The aim of this project is to understand the effect of background noise on the processes underlying non-native spoken-word recognition. In this presentation, I will present recent results on the effect of background noise on 1) the flexibility of the perceptual system in non-native listening; 2) the multiple activation, competition and recognition processes in non-native spoken-word recognition.

Biography: Odette Scharenborg (PhD) is an associate professor at the Centre for Language Studies and a research fellow at the Donders Institute for Brain, Cognition, and Behaviour, Radboud University (Nijmegen, the Netherlands). Her research interests focus on narrowing the gap between automatic and human word recognition. In 2008, she co-organized the Interspeech 2008 Consonant Challenge, which aimed at promoting comparisons of human and machine speech recognition in noise in order to investigate where the human advantage in word recognition originates. She was one of the initiators of the EU funded Marie Curie Initial Training Network 'Investigating Speech Processing In Realistic Environments' (INSPIRE, 2012-2015). Her current project is funded by a fellowship from the Netherlands Organisation for Scientific Research on
the topic of human non-native word recognition in noise, which will be investigated using a combination of listening experiments and computational modelling.

Host: Prof. Shrikanth Narayanan & Prof. Panayiotis Georgiou

Location: Hughes Aircraft Electrical Engineering Center (EEB) - 132

Audiences: Everyone Is Invited

Contact: Tanya Acevedo-Lam/EE-Systems