Learning Mechanism: Function Approximation

LMS Weight update rule:

Do repeatedly:

• Select a training example b at random

  1.

  Compute error(b):
  

  $$error(b) = V_{train}(b) - \hat{V}(b)$$

  
  

  2.

  For each board feature f_i, update weight w_i:
  

  $$w_{i} \leftarrow w_{i} + c \cdot f_{i} \cdot error(b)$$

  
  

c is some small constant, say 0.1, to moderate the rate of learning