Awesome  explanation Fergal :)

So, for 128 bit scalar encoding atleast required 500 columns to see the
stable SDR, sequence learning. For example, if i have large (i.e 10k) bit
length of encoded scalar values, 2048 columns enough to get the above
effects?

Thanks
Ramesh Ganesan



On Fri, Aug 9, 2013 at 6:20 AM, Doug King <[email protected]> wrote:

> Very nice, concise description of the Spatial Pooler. Thanks Fergal.
>
>
> On Wed, Aug 7, 2013 at 8:32 AM, Fergal Byrne 
> <[email protected]>wrote:
>
>>
>> Hi Francisco,
>>
>>  The Spatial Pooler is an artefact of NuPIC; it simulates the way a
>> layer of columns becomes activated by feedforward inputs. In essence, the
>> algorithm is:
>>
>>  Get all columns to check their feedforward dendrite and generate an
>> activation score.
>> Pick the top 2% of columns based on activation score and activate them.
>> Update the permanences on the feedforward dendrites.
>>
>>  In the neocortex, the columns do this themselves, and they pick the top
>> 2% (may vary) by higher-score columns inhibiting nearby lower-score
>> columns. They also update their own synapses themselves.
>>
>>  So the SP is simulating the inter-region communication as well as the
>> single-region sensory communication.
>>
>>  Regards,
>>
>>  Fergal Byrne
>>
>>
>> _______________________________________________
>> nupic mailing list
>> [email protected]
>> http://lists.numenta.org/mailman/listinfo/nupic_lists.numenta.org
>>
>>
>
> _______________________________________________
> nupic mailing list
> [email protected]
> http://lists.numenta.org/mailman/listinfo/nupic_lists.numenta.org
>
>
_______________________________________________
nupic mailing list
[email protected]
http://lists.numenta.org/mailman/listinfo/nupic_lists.numenta.org

Reply via email to