[mlpack] contribute to RBFN

Marcus Edel marcus.edel at fu-berlin.de
Mon Mar 2 13:23:40 EST 2015


Hello Shangtong,

Sorry for the slow response. Great to see that you are interested in
implementing Radial basis function networks.

>   […] I think it’s better to model the whole RBFN as one layer class in current framework of ANN.
> Then this RBFN layer can be used to build other network such as CNN, FFNN.
> Is it appropriate? […]

Sounds like a great idea, in fact my plan is to use the network for some
reinforcement learning.

Thanks,
Marcus

P.S.: I need some more time to go through the rest of the convolution neural
network code you're coding output and adjustment is way faster than my ability
to check the code.


> On 02 Mar 2015, at 07:35, Shangtong Zhang <zhangshangtong.cpp at qq.com> wrote:
> 
> Hi,
> 
> I want to contribute to RBFN.
> Here is my understanding about implementing RBFN in current framework.
> According to wikipedia,  http://en.wikipedia.org/wiki/Radial_basis_function_network <http://en.wikipedia.org/wiki/Radial_basis_function_network>,
> RBFN has three layers, 
> but I think it’s better to model the whole RBFN as one layer class in current framework of ANN.
> Then this RBFN layer can be used to build other network such as CNN, FFNN.
> Is it appropriate?
> 
> Best
> Regards
> 
> 
> Shangtong Zhang,
> Third Year Undergraduate,
> School of Computer Science,
> Fudan University, China.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.cc.gatech.edu/pipermail/mlpack/attachments/20150302/ddf4cfaa/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5136 bytes
Desc: not available
URL: <https://mailman.cc.gatech.edu/pipermail/mlpack/attachments/20150302/ddf4cfaa/attachment.bin>


More information about the mlpack mailing list