Computational modeling of early language acquisition – Bridging the gap between acoustic input and high-level linguistic representations
Date:
Fri, 06/19/2015 - 11:00am - 12:30pm
Location:
CCRMA Seminar Room
Event Type:
Hearing Seminar Who: Okko Räsänen (Stanford and Aalto University)
What: Computational modeling of early language acquisition
Bridging the gap between acoustic input and high-level linguistic representations
When: 11AM on Friday June 19, 2015
Where: CCRMA Seminar Room
Why: Because understanding sound is hard for us, how do infants do it????
Title: Computational modeling of early language acquisition
Bridging the gap between acoustic input and high-level linguistic representations
Okko Räsänen (Stanford and Aalto University)
Children learn their native language relatively effortlessly while simply interacting with their language community. This is an extraordinary achievement when we start to consider the challenges involved in early language acquisition. For instance, starting without any a priori linguistic knowledge, infants have to learn to segment words out of continuous speech, how these words map to their referential meanings, and how words are made up of smaller units such as phones and syllables that do not carry any meaning in isolation. Despite our growing scientific knowledge of early language acquisition, there is still a significant gap between the speech representations that are known to be accessible to early language learners from the sensory input and the linguistically defined characterizations of language that we often take as the ground truth for language structure. In this presentation, I will give an overview of the early language acquisition problem from an infant’s point of view and describe how this problem can be approached using computational modeling, i.e., speech signal processing and machine learning. I will also attempt to draw some parallels between machine learning and the so-called statistical learning paradigm that has become very influential in the study of human learning and cognition.
Bio
Okko Räsänen was born in Finland in 1984. He received the M.Sc. degree in language technology from the Helsinki University of Technology in 2007 and D.Sc. (Tech.) degree in language technology from Aalto University, Finland, in 2013. He is currently a postdoctoral researcher at the Department of Signal Processing and Acoustics at Aalto University and a visiting researcher at the Language and Cognition Lab of Stanford University. His research interests include computational modeling of language acquisition, cognitive aspects of language processing, context-aware computing, multimodal data analysis, and speech processing and psychoacoustics in general. He is a member of ISCA and Cognitive Science Society.
FREE
Open to the Public