
The system works by teaching the computer to assess large amounts of data and screening for patterns. It then takes these patterns and compares it to patterns shown by people with communication disorders. The technology here needs to be able to listen to the child and recognize their voice, and also analyze the pattens shown. Specifically, the computer listens to the pauses in speech.
I like to compare this machine's ability to scan large amounts of information to IBM Watson and how it is also being used in the medical field for cancer patients. In addition, it uses "machine learning" which is also present in IBM Watson. This is the ability of a computer to "learn without being explicitly programmed" (Wikipedia).
Although there are still some bugs in the system and problems that need to be sorted out and accounted for, this technology is wildly promising. Detecting communication disorders early is crucial in allowing for academic success later in life. Hopefully this technology is perfected soon, because it would be so beneficial for future generations!
________________________________________________________________________________
https://en.wikipedia.org/wiki/Machine_learning
http://news.mit.edu/2016/automated-screening-childhood-communication-disorders-0922
https://www.southernct.edu/academics/schools/health/academic-programs/communicationdisorders/centerforcommunicationdisorders/
This is a very important use for computers, not just in the medical field: simply processing more data with a greater sensitivity for subtlety than we can achieve ourselves. There may be people who were never diagnosed because we're simply not good enough at detecting the problems ourselves.
ReplyDeleteIt's interesting how - regarding mental disorders - we're the only ones who can determine whether one has a certain condition. Not some blood test that has a more-definite answer, just our skills of observation. Putting a task like this on a computer would be greatly beneficial for diagnosis to be more accurate than ourselves. There are many tasks we've already given to computers to lessen or even eliminate human error and spotting disorders such as these should make a great candidate.