
Ai startup Companies engaging in data prepration are making money
ML - The way the world works - analyzing how things work
00:00
Is the a I Better Than a Professional Radiologist?
Image net had 23 layers, and then it pulled those features from those layers. Then it had a dense layer at the end for its labelling which was discreet. And thata then allowed it to perform very well, in fact, better than most humans could do,. That's because it had more exposure to more images types. However, this does bring up a good point about domain knowledge. If you take, for example, a a and look how well it does at recognizing radiology images for cancerTake or tumor patients; there's a lot of false alarms or false negatives that its generating.
Play episode from 03:08
Transcript


