Hi,

Hans wrote:

>>These applications could be implemented as Csound opcodes or perhaps a
more
>>flexible neural network implementation could be developed.  Maybe a
>>perceptron or neural network opcode could be implemented with weights
>>supplied in table form so different functions could be implemented with a
>>single opcode.


Pedro wrote:

>a neural ugen would have to be
>fully flexible in structure, and data independent, so that it could be used
>to learn about anything, allowing to implement all the above applications;
>problem is still the slowness of the back-prop algo; some faster algorithms
>exist, and when coding in C some reduction to integer math could help
>speeding things up;

I think that perhaps a utility program would be useful for training the net.
Perhaps there is a popular, publicly available, cross-platform ANN package
which already exists.  If the Csound opcode were able to read this file
format you could do the training in this "utility" program and then use the
result in Csound.  Then you would not have to reinvent the wheel so much.

I realize that most of the neural applications I mentioned already exist but
they are not available in the Csound environment.

Bye,
Hans Mikelson