(cont'd from last Friday)
Dr. Lennox explains "narrow" artificial intelligence (AI) with this practical example of its application in health care.
A person with a diseased lung goes to a doctor for a diagnosis. An x-ray of the lung is taken; the doctor looks at it and compares the x-ray to his/her memories of articles and other x-rays he/she has seen. Using his/her memory and knowledge, a diagnosis is made.
Or . . the patient's x-ray is fed into an AI system which compares it to the million lung x-rays of other patients stored in it. It looks for patterns, and a diagnosis is made without delay.
Which method is faster and more accurate? Probably the AI method, provided that correct information and meaning of the x-rays has been programmed into it by wise doctors.
Narrow AI usually does just a single thing which normally would have required natural (human) intelligence. But it doesn't know what it's doing. It is simply programmed to do a methodical task using a large amount of input.
(cont'd tomorrow)
No comments:
Post a Comment