Uh - WOW!
First - everyone should watch this video now. Chang is very bright, he gives a spontaneous, clear, deep evaluation of AI in radiology in 7 minutes.
Highlights - A lot of hype, which will not come to pass in the near future, but because it is sexy, gets the popular press all excited, which, not incidentally, drives money into these start ups.
Yeah.
Where AI will have some effect in the near future: mundane things like improving our acquisition protocols.
Current IT infrastructure is insufficient to handle much in the way of AI. Infrastructure needs to be improved.
And then - he mentions, pointedly, that some data scientists put pre prints out on arXive that have very poor clinical components of the research, but nevertheless are picked up by the less sophisticated press and trumpeted to the skies. Brian Casey says "I think I know which paper you are talking about" So do I. ;)
For Brian C - what a dream interview, you ask two questions, and he talks for 7 minutes giving a very listenable, very meaty presentation off the top of his head. I don't think I have ever seen anything quite like this.
I agree entirely, a fantastic summary indeed.
I occasionally check the twitter feed and the hype is really amazing - and somewhat amusing, seeing all this bandwagoning, and people with no domain specific knowledge suddenly reiterating clichés ad nauseam...
Just watched it.
Pretty good summary.
I'll give it a +1 :)
Seemed kind of nervous about it to me, but I may be alone in thinking this.