Hi Matt, Just to clarify: in this video:
https://www.youtube.com/watch?v=5r1vZ1ymrQE&feature=youtu.be&t=8m16s Was Nupic predicting the entire tune without input or was it just predicting the next note given the current note? And is the source code for this example available? On Tue, Jun 30, 2015 at 5:33 PM, Matthew Taylor <[email protected]> wrote: > On Tue, Jun 30, 2015 at 8:06 AM, Julian Samaroo <[email protected]> wrote: >> I could, right now, code a simple TD-learning system (which is >> well-described in literature), to perform actions, and then I could slap on >> a convolutional neural network to allow it to behave in response to certain >> images of people, places, objects, or really whatever I'd like. Why is >> Numenta so slow to add something so simple? These things have been done >> time-and-time again, just like HTM even. > > There is nothing stopping you from doing this. I would be extremely > interested in someone building something like this. It is a very cool > idea. Why don't you do it? > > --------- > Matt Taylor > OS Community Flag-Bearer > Numenta >
