Code & Fun


  • humman learning
    • acquiring skill with experience accumulated from observation
      即 observation -> learning -> skill
  • machine learning:
    • auquiring skill with experience accumulated/computed from data

skill <=> improve some performance measure
Machine Learning: improving some performance measure with experience computed from data.

The key essence of Machine Learning:

  1. exists some underlying pattern can be improved
  2. but no programmable (easy) definition
  3. somehow there is data about pattern.


  • Input: $x \in X$
  • Output: $y \in Y$
  • target-func: $X->Y$

the target-func <=> unknown pattern to be learned.

data <=> traning example <=> $D = { (x_1,y_1),(x_2,y_2),… }$

hypothesis <=> skill with hopefully good example.


Machine Learning:

use data to compute hypothesis g that approximates target f.

本文首发于Code & Fun