Massachusetts General Hospital's Institute of Health has been working with researchers from the Computer Science department at MIT to create a computer system that automatically determines whether or not a child has a speech or language disorder. It's important to diagnose these disorders at a young age so the children can learn to grow out of the disorder by the time their an adolescent. Unfortunately, sixty percent of children go undiagnosed by the time they reach kindergarten. This system works to diagnose speech and language disorders by analyzing children's audio performances on reading a story. The children watch a series of images and narrative about a story and then they need to tell the story back in their own words. To check how accurate the system was researchers had to, "use a standard measure called area under the curve, which describes the tradeoff between exhaustively identifying members of a population who have a particular disorder, and limiting false positives"(Hardesty). The researchers' performed three tests to find its accurate about eighty percent of the time. In medicine, if the system works more than seventy percent of the time it is considered an accurate test.
Two graduates of MIT, John Guttag and Jen Gong, believed that pauses in children's speech, when they try to complete sentence or remember a word, are sources that help diagnose communication disorders. So they implemented thirteen acoustic features of children's speech into their system to be recognized. Their system recognizes certain patterns of pauses and error in speech that correlate to the communication disorders it can diagnose. Some of the acoustic features it can recognize are length of pauses, short or long pauses, and variability of the length of the pauses. Thomas Campbell, a professor of behavioral and brain sciences at the University of Texas at Dallas says, "The researchers’ automated approach to screening provides an exciting technological advancement that could prove to be a breakthrough in speech and language screening of thousands of young children across the United States"(Hardesty).
Two graduates of MIT, John Guttag and Jen Gong, believed that pauses in children's speech, when they try to complete sentence or remember a word, are sources that help diagnose communication disorders. So they implemented thirteen acoustic features of children's speech into their system to be recognized. Their system recognizes certain patterns of pauses and error in speech that correlate to the communication disorders it can diagnose. Some of the acoustic features it can recognize are length of pauses, short or long pauses, and variability of the length of the pauses. Thomas Campbell, a professor of behavioral and brain sciences at the University of Texas at Dallas says, "The researchers’ automated approach to screening provides an exciting technological advancement that could prove to be a breakthrough in speech and language screening of thousands of young children across the United States"(Hardesty).
Resources:
Hardesty, Larry. http://news.mit.edu/2016/automated-screening-childhood-communication-disorders-0922
https://techcrunch.com/2016/09/23/machine-learning-could-automate-screening-kids-for-speech- and-language-disorders/
No comments:
Post a Comment