Google Patent Talks About Advantages of Using a Machine-Learned Model Instead of a Information Retrieval Based Model

Bill Slawski breaks down a newly granted Google patent Searchable index that sheds some light on how Google may use a machine-learned model to produce more accurate recommendations in search.

The patent tells us about the difference between the two:

• An information retrieval tool is computationally efficient, but can only produce a rough estimate of which items are best recommended to a user.
• A machine-learned model can produce more accurate recommendations but is often more computationally intensive than an information retrieval tool.

Check out Bill’s post for the full breakdown.