Hello, Glad you found that useful. One thing I didn't mention that you brought up was the difference between flavour and smell. In food, what we normally describe as the flavour is mostly the smell [1]. When you eat food the odors work their way up the back of your throat into your nose. That gives you the majority of the flavour of food. Your tongue only contributes a few tastes (sweet salt, butter, sour, etc). I think for that reason databases only list the odor compounds in food. Beyond smell and taste there are lots of other factors that effect flavour. For example texture, mouth-feel, color and visual appearance.
It would be cool to be able to incorporate all these things into the CLA . The more information you can give it the more patterns it can pick out. But I'm not sure of the best way to do this, but maybe others can chip in. Also I found a few of the databases give you limited access for free. If you go to [2] and click sign as guest you can explore different foods. [1] - http://en.wikipedia.org/wiki/Flavor [2] - http://www.vcf-online.nl/VcfProducts.cfm Cheers, Ruaridh On Sat, Apr 12, 2014 at 12:03 AM, Chetan Surpur <[email protected]> wrote: > > On Fri, Apr 11, 2014 at 12:35 PM, Matt Roesener <[email protected]>wrote: > >> Is there a work around this? > > > Requesting clarification: a workaround for what? Lack of multi-step > prediction when using CEPT + temporal pooler? > > _______________________________________________ > 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
