y... very cool... and useful! On Tue, Feb 16, 2016 at 12:19 PM, Matthew Taylor <[email protected]> wrote:
> Marcus, that's really cool. I love using d3. I'll bet you did a lot of > work optimizing those charts! It is really neat to see the affects of > bursting. > --------- > Matt Taylor > OS Community Flag-Bearer > Numenta > > > On Mon, Feb 15, 2016 at 8:02 PM, Chandan Maruthi > <[email protected]> wrote: > > Marcus, > > This is awesome. And a well written post . You should get this on the > wiki. > > It helps people understand what Nupic can do before committing too much > > time. That should get more people > > to explore the possibilities further with HTMs. > > > > Regards > > Chandan > > > > On Mon, Feb 15, 2016 at 6:12 PM, Marcus Lewis <[email protected]> wrote: > >> > >> Hey all, > >> > >> I just posted: See your HTM run: Stacks of time series > >> > >> Late last year Chetan said: > >>> > >>> It would also be useful to get plots for the other data in the Monitor > >>> Mixins, like # of segments, # of connected synapses per segment, etc., > to > >>> get a high-level picture of how the TM is learning. > >> > >> > >> I ran with the idea. It's been my side project for a few weeks. > >> > >> I built this with d3. It's not bundled into Sanity because it seems > useful > >> to start out unbundled. If you ever want to make something like this, I > >> recommend snatching my code or at least using it as a reference. > >> Visualizations of 4000 timesteps are not performant by default. > >> > >> This has been fun because I didn't know what story the data would tell. > >> For example, the somewhat-linear growth of segments (spikable and > >> connectionless) with hotgym. I've spent a lot of time at the hotgym and > I > >> didn't know it'd look like that. > >> > >> Feedback welcomed! > >> Marcus > > > > > > > > > > -- > > Regards > > Chandan Maruthi > > > >
