The Back-propagation algorithm

The back-prop algorithm for supervised learning from exemplars is:

Repeat:

1. Send in inputs I_i for this exemplar.
2. Calculate outputs y_k
3. Get given correct outputs O_k
4. Measure E.

5. w_jk weights:
   1. Calculate all the   =   y_k ( 1 - y_k ) ( y_k - O_k )
   2. Calculate all the   =     y_j
   3. For all j,k:
      

6. w_ij weights:
   1. Calculate all the   =   y_j ( 1 - y_j )   Σ _k ( w_jk )
   2. Calculate all the   =   I_i
   3. For all i,j:
      

7. Repeat (next exemplar).

---

See sample code.