This Whiteboard Friday, Ben suggests some of the things the search engines do in terms of the overall trends of the machine learning community.
Before Google, there were many methods of machine learning. One of the things Google says they’ve done is look at the page and query, in conjunction with multiple other unknown signals.
There are a couple of exciting things going on in the literature regarding this, with these being a few of the movements:
Disjoint algorithms to joint algorithms
Shallow algorithms to deep algorithms
Discriminative models to generative models