Last modified: Wednesday, September 19, 2007
$1 million grant to fund toddler word-learning study
FOR IMMEDIATE RELEASE
Sept. 19, 2007
BLOOMINGTON, Ind. -- Researchers in Indiana University Bloomington's Department of Psychological and Brain Sciences have received a $1 million grant from the National Institutes of Health to study how the brain uses highly complex statistics to learn language.
The study, led by assistant professor Chen Yu and Linda B. Smith, professor and chair of the department, will involve an interdisciplinary research team with expertise in both behavioral studies and computer science.
The research began with the discovery that toddlers, when they begin to learn words, can simultaneously and rapidly learn many word-object pairings by internally computing complex statistics. The team will use advanced sensing equipment, such as an eye tracker, to study learning processes and to develop computational models and systems to understand this learning.
"This new discovery changes completely how we understand children's word learning, which has been thought of as happening one word at a time, rather than as a kind of data-mining of all the ambiguous word-referent pairings babies happen to hear," said Smith.
Results from behavioral studies and computational modeling suggest that human learning mechanisms are complex, adaptive, extensive and available to babies, said Yu.
"The next step is to explore what kinds of computational and cognitive abilities babies and people more generally bring to the learning process," he said.
The research project began a year ago and originally was funded by IU's Faculty Research Support Program. The NIH grant will allow the study to continue for another five years.
Research for this study is primarily conducted in the department's Computational Cognition and Learning Lab, which Yu directs. The goal of the lab's research is to understand human development and learning through both behavioral studies and computational modeling.