Closed byzhang closed 10 years ago
Hm, I've thought about this a bit and just don't think it will fit into the framework that the theanets package establishes. You could definitely write an implementation of PSD using theano, but there are several important differences between a standard feedforward network and how PSD works (especially the alternating block learning procedure), so I think it would be difficult to include.
Thanks for commenting on it.
@lmjohns3 do you happen to know other framework can support PSD?
I'm new to theano and theano-nets, so sorry if it's a trivial question. But I'm wondering could you add an example for predictive sparse decomposition as in the paper http://cs.nyu.edu/~koray/publis/koray-psd-08.pdf?