Take Healthcare. We can assume IBM has access to lots of individual health records, as well as all the scientific literature about - I guess - the corresponding health problems. Suppose they set about pairing the phrases used in the literature with examples of individual symptoms?
With a good understanding of the narrative patterns in the scientific literature, you might say that the correlation engine's job would be to pair stories to symptoms, and come at the narrative of diagnosis in a different way, through the back door.
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment