Hi Marcus Thanks a lot for the reply and also guiding me on the IRC(my nick is sagarbhathwar). I am researching on ssRBM and once I am done with thoroughly understanding it, I will finish up the proposal for ssRBM. Shall I put it for feedback/review as soon as I complete ssRBM(+DBN) or should I also complete the proposal for Stacked GAN(which I think is our way forward for GANs) and then put it up for review together? I shall follow according to your convinience
P.S. - Is it a good idea to separate the project details into two parts? First part - "The high level abstraction details" which talks about the algorithm in abstracted terms without going into coding the algorithm Second part - "The low level implementation details" which talks about how the code for the algorithm adhers to mlpack API and can interact with existing API to make full utilization of it. What do you think about the above approach? Thanks a lot for your time Sagar B Hathwar On Sun, Mar 26, 2017 at 2:47 PM, Marcus Edel <[email protected]> wrote: > Hello Sagar, > > welcome and thanks for the interest. > > > I titled the project as "Generative Neural Networks - RBM and GAN". Any > thoughts > > on this? > > Sounds good, note that the focus of the project is to take a look at recent > improvements/ideas, meaning that implementing e.g. standard RBM might be an > option but it should be implemented with the extensions in mind. > > > Can I also have some pointers on writing a proposal on this topic? I went > > through the guidelines on the wiki, but since this project involves > implementing > > two algorithms, I would like to know the best way to go about writing a > good > > proposal > > There are a lot of good proposals out there that might be helpful, the > Student > Manual (https://developers.google.com/open-source/gsoc/resources/manual) > also > links to some good examples. > > I hope this is helpful, let us know if you have any further questions. > > Thanks > Marcus > > > On 25 Mar 2017, at 20:16, Sagar B Hathwar <[email protected]> > wrote: > > > > Hello > > > > I looked at the Essential Deep Learning Modules under GSoC project > ideas. Since generative models are a hot field today, I felt that > implementing Restricted Boltzmann Machine(plus Deep Belief Networks) and > Generative Adversarial Networks (for unsupervised learning) would be a > good combination since both are generative but quite different. I went > through the resources provided under the topic and also additional papers > on it. Both these algorithms are very interesting. > > > > I titled the project as "Generative Neural Networks - RBM and GAN". Any > thoughts on this? > > > > Can I also have some pointers on writing a proposal on this topic? I > went through the guidelines on the wiki, but since this project involves > implementing two algorithms, I would like to know the best way to go about > writing a good proposal > > > > Thanks > > Sagar > > _______________________________________________ > > mlpack mailing list > > [email protected] > > http://knife.lugatgt.org/cgi-bin/mailman/listinfo/mlpack > >
_______________________________________________ mlpack mailing list [email protected] http://knife.lugatgt.org/cgi-bin/mailman/listinfo/mlpack
