On Sat, Nov 23, 2013 at 12:25 AM, Scott Purdy <sc...@numenta.org> wrote:
> > For some reason having trouble wrapping my head around that but I believe > that all makes sense. I think you are saying that if your input space was > varied enough that you saturate the SP, then the number of patterns you can > represent is inversely proportional to the noise tolerance. The caveat to > this is that even if you end up with a different SP representation because > of a small amount of noise, the representation will still be very > semantically similar to the previous. In fact, you most likely only have > one or two columns that are different so the higher levels will see a very > similar pattern. > Yes, I wouldn't say it better ;) Thanks for rewording to make sense. (I was just talking noise at the output layer, where 10% noise would make 100 from a 1000 bits, that could easily change meaning if you're running on 20 ON bits). Btw, could this be an argument for reconstruction against classification? When I mess with 10% of the SDR, you'll train on a wrong pattern, if we use "back-propagation" the random-bits will have low permanences, so will reduce on the way down. > >> >> PS: >> Is there a (lower bound) limit on the number of columns in SP? So would a >> 20 col SP work? That way, I could achieve the (20 choose 3) and reach the >> state of info-full SP. >> > > The theory relies on large numbers. Subutai's CLA quiz covers it very > thoroughly. In your 20 choose 3 example, you lose fault > tolerance/subsampling at higher levels, the ability to represent many > different patterns (only 1140), and the ability to represent many > simultaneous patterns. > I know, the fault tolerance would be a problem. Lost ability to represent huge num. of patterns is what I want. Otherwise it's out of reach for anybody to experiment with a SP in nearly saturated state (input pattern wise). -- Marek Otahal :o)
_______________________________________________ nupic mailing list nupic@lists.numenta.org http://lists.numenta.org/mailman/listinfo/nupic_lists.numenta.org
