Next: 5.3 Experiments
Up: 5.2 Pattern association algorithm
Previous: 5.2.2.1 Examples
5.2.3 Recall
In recall, the components of a pattern are partially given, and the components of a completed pattern are divided into input and output (see section 4.2). The input is the offset of a hyperplane extending into the output dimensions:
where
are the free parameters of the hyperplane. The matrix defines which dimensions are input and which are output. To obtain the output for a given input, the free parameters are chosen such that the potential
E() is minimized,
^{*} 
= 
argE(()) , 


= 
^{*} + . 
(5.11) 
There seems to be no analytical solution for
^{*}. In feature space, the constraint space is not a hyperplane. However, the minimization can be solved using standard numerical optimization algorithms.
Next: 5.3 Experiments
Up: 5.2 Pattern association algorithm
Previous: 5.2.2.1 Examples
Heiko Hoffmann
20050322