If we had reconstruction we could use the SP and CLA classifier together to get multi step predictions, and reconstruct them to the CEPT SDR space. On Apr 12, 2014 11:36 AM, "Matt" <[email protected]> wrote:
> Hi, Yes a work around to creating multistep predictions using CEPT. I > think I understand the problem since CEPT is creating the SDRs. But to use > the CLA classifier will I have to create my own SDRs of each ingredients? > One thing that was brought up earlier about this, was to somehow append the > unique chemical compounds or odorants that make food taste the way it > tastes. There are hundreds and thousands of variations that might go well > together. So creating new SDRs for specific ingredients instead of using > CEPT might be a better idea. > > Thanks for your help I appreciate your insight. > > Matt > > Sent from my iPhone > > > On Apr 12, 2014, at 9:00 AM, [email protected] wrote: > > > > Re: Food Classification/ Ingredient Prediction > > _______________________________________________ > 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
