Skip to content

Latest commit

 

History

History
10 lines (5 loc) · 397 Bytes

TransformLearning.md

File metadata and controls

10 lines (5 loc) · 397 Bytes

#DataScience

There is usual #problem on how to apply #AI,#ML-engine to Data2, when it was trained on Data1.

The method is called #transformlearning and think of it like this:

  • Trained engine is an operator, which maps Data1 to wanted Outcome: E1(Data1)=Outcome1

  • If Data2 is "similar" to Data1 (think of it like "distance" within "similarity metrics"), then one can do E1(Data2)=Outcome2