2013/9/16 Issam issamo...@gmail.com:
Maybe I could start working on the simple approach - 'Gaussian visible
nodes'.
There are several variants of the RBM, for various types of input:
* Bernoulli RBM for binary inputs (Bernoulli random variables; this is
what we have)
* Gaussian RBM for
And under the current implementation, implementing them involves
changing only the sampling and energy computation, I think.
I discussed this with Gabriel Synnaeve during the sprint and I think
he was working on the gaussian version, it might be on his repo.
Lars, do you have any practical
2013/9/17 Vlad Niculae zephy...@gmail.com:
And under the current implementation, implementing them involves
changing only the sampling and energy computation, I think.
I discussed this with Gabriel Synnaeve during the sprint and I think
he was working on the gaussian version, it might be on
Thanks guys!
--Issam
On 9/17/2013 12:30 PM, Olivier Grisel wrote:
2013/9/17 Vlad Niculae zephy...@gmail.com:
And under the current implementation, implementing them involves
changing only the sampling and energy computation, I think.
I discussed this with Gabriel Synnaeve during the sprint
Reading the documentation, it seems that scikit's RBM does not support
continuous values or, more precisely, values that are larger than one or
less than zero; since logistic is the output function.
This paper, http://www.ee.nthu.edu.tw/~hchen/pubs/iee2003.pdf
On 09/16/2013 09:54 PM, Issam wrote:
Reading the documentation, it seems that scikit's RBM does not support
continuous values or, more precisely, values that are larger than one
or less than zero; since logistic is the output function.
This paper,
Good idea.
Section 13.2 in this tutorial
http://www.cs.toronto.edu/~hinton/absps/guideTR.pdf
http://www.cs.toronto.edu/%7Ehinton/absps/guideTR.pdf , explains some
standard techniques on dealing with real-valued visible nodes.
I also found a great answer here:
