New Thoughts On Language Acquisition: Toddlers As Data Miners
Indiana University researchers are studying a ground-breaking theory that young children are able to learn large groups of words rapidly by data-mining.
Their theory, which they have explored with 12- and 14-month-olds, takes a radically different approach to the accepted view that young children learn words one at a time — something they do remarkably well by the age of 2 but not so well before that.
Data mining, usually computer-assisted, involves analyzing and sorting through massive amounts of raw data to find relationships, correlations and ultimately useful information.
[…]
“This new discovery changes completely how we understand children’s word learning,” Smith said. “It’s very exciting.”
Wow, that’s fascinating. I coded a variation of Han’s FP-tree about a year ago. It’s odd to think of it working like that in the human mind.
“The learning mechanisms used by the children to learn words also could be used to further machine learning.”
Simply amazing. Maybe we can actually get a machine to learn a few words and get that ‘voice to text’ feature Heinlein et al. promised us in the 1950s.
In best Jimmy Doohan (Scotty) voice:
“Hello, computer … How quaint, a keyboard.”
Isaac Newton couldn’t speak until he was four.
A developed nation’s human rights violation….
children working in data mines.