Hi David, It's not just the "mainstream", the Deep Learning people and most of these AGI guys also have a low opinion of the whole HTM program. The Deep Learning guys were being laughed at for over a decade until their software started winning all the benchmark competitions (and now they dominate them), so they are in turn mocking us and demanding we publish papers and compete in their competitions. They regard anything we say until then as "marketing BS" and say the only reason HTM is being researched is because Jeff has money and kudos from his Palm days.
I've countered (most recently on reddit) that we are primarily doing brain science, and that the machine learning is a fortunate by-product which is not yet ready to compete with Deep Learning on their (recently conquered) turf. I get some intelligent conversations but mostly just the same old dismissive replies. What they say, and perhaps it's worth thinking about, is that we would gain credibility if we could demonstrate a "killer application" (what Ben Goertzel calls the "AGI Sputnik moment") of HTM which shows it solving a problem nobody else can even attempt to solve. For that, I believe, we'll need the full sensorimotor stack and hierarchy, which we will have in the next few months. Could I ask people to have a think about this and possibly bounce around ideas? We could schedule a round-table session during the hackathon next month and see if there's an application area to focus on in this regard. The most immediate candidates I can see right now are cortical.io (aka CEPT) for NLP and the Geospatial Encoder. Regards Fergal Byrne On Mon, Sep 29, 2014 at 2:10 PM, cogmission1 . <[email protected]> wrote: > From reading this, it seems a lot of people are forming opinions out of > past skepticism and disbelief. It's also apparent that these opinions are > formed without investigation - so there may be something to overcome in the > main stream. > > I think, once the sensor-motor stuff is complete and installation is > fire-and-forget, there should be a rather largish marketing campaign to get > things in the "proper" perspective. People *need* to know how uniquely > outstanding the work at Numenta is! > > David > > On Sun, Sep 28, 2014 at 5:44 AM, Richard Crowder <[email protected]> wrote: > >> Along with this interview of Dr. Itamar Arel a few years back (interview >> by Ben Goertzel); >> >> http://hplusmagazine.com/2011/02/04/itamar-arel-on-the-path-to-artificial-general-intelligence/ >> >> >> On Thu, Jul 24, 2014 at 3:05 AM, Richard Crowder <[email protected]> >> wrote: >> >>> And the HTM is mentioned in this ahead of print overview of AGI; >>> >>> http://www.degruyter.com/printahead/j/jagi >>> >>> >>> >>> On Mon, Jul 21, 2014 at 4:56 AM, Chetan Surpur <[email protected]> >>> wrote: >>> >>>> A link to NuPIC is on the front page of Hacker News right now (with 10 >>>> points): >>>> >>>> https://news.ycombinator.com/item?id=8062175 >>>> >>>> >>>> >>>> _______________________________________________ >>>> nupic mailing list >>>> [email protected] >>>> http://lists.numenta.org/mailman/listinfo/nupic_lists.numenta.org >>>> >>>> >>> >> > > > -- > *We find it hard to hear what another is saying because of how loudly "who > one is", speaks...* > -- Fergal Byrne, Brenter IT http://inbits.com - Better Living through Thoughtful Technology http://ie.linkedin.com/in/fergbyrne/ - https://github.com/fergalbyrne Founder of Clortex: HTM in Clojure - https://github.com/nupic-community/clortex Author, Real Machine Intelligence with Clortex and NuPIC Read for free or buy the book at https://leanpub.com/realsmartmachines Speaking on Clortex and HTM/CLA at euroClojure Krakow, June 2014: http://euroclojure.com/2014/ and at LambdaJam Chicago, July 2014: http://www.lambdajam.com e:[email protected] t:+353 83 4214179 Join the quest for Machine Intelligence at http://numenta.org Formerly of Adnet [email protected] http://www.adnet.ie
