Does a pulsed (spiking) neural network outperform a recurrent neural net?

Question

Both pulsed and recurrent neural networks can model time-varying information. But I am not sure which model is better relative to the computational cost. Does it pay to use the more complex pulsed neural network or will the recurrent neural network just as well with a lot less calculation required? Does a pulsed net converge more quickly?

Thanks

Answer 1

The most important property of a neural network (or any auspicious model) is its accuracy. I would rather spend 10 times longer building the model if it were significantly more accurate (and auspicious).

There are several standard techniques for assessing the predictive power of your model, such as

•  leave-one-out cross-validation

•  leave-many-out cross-validation

•  Fisher randomization (http://en.wikipedia.org/wiki/Ronald_Fisher)

There are also many guiding principles for building a predictive model, such as

•  Occam's razor

•  avoid overfitting (http://web.engr.oregonstate.edu/~tgd/classes/534/slides/part10.pdf)

•  penalties for overfitting (http://en.wikipedia.org/wiki/Regularization_(mathematics))

Here are a few areas to look for more information

http://predictivemodelingblog.blogspot.com/

http://www.statsoft.com/textbook/data-mining-techniques/