Mark, You've probably seen this, but just in case you haven't : http://inbits.com/2014/09/a-unifying-view-of-deep-networks-and-hierarchical-temporal-memory/
On Thu, Oct 8, 2015 at 7:10 AM, Marek Otahal <[email protected]> wrote: > Hi Matt, David, > > thanks for your updates. > > On Thu, Oct 8, 2015 at 1:45 PM, cogmission (David Ray) < > [email protected]> wrote: > >> Hi Matt, >> >> As always, I am inspired by your community leadership and endless >> support. The way in which you nurture community engagement is as much a >> phenomenon as the HTM technology itself. You have made such a difference in >> my life personally (I can't imagine where I'd be right now if it weren't >> for that fateful phone call 1 year ago!). >> >> I think "What is Riverview?" >> > Yup! Streaming data toolkit, aka Riverview is pretty cool. I plan to > review it for some HTM Challenge ideas. > > >> is a very interesting question and a metaphor for your unrelenting >> support of the HTM endeavor. I think it is subtly one of the most >> significant developments in support of HTM. A central nexus and repository >> for time based streaming data, and a general purpose temporal data >> collection and composition service. Sound good? I think it is awesome just >> how subtly and without fanfare Riverview was conceived - hiding its >> enormous significance. It's kinda funny, that way... >> >> Anyway, here's my list: >> >> 1. Readying myself (learning as much as possible) so that I can assist >> with HTM & Cortical.io/Semantic Folding integration. >> > This is interesting, could you provide some more insight David? > > > I'm currently: > > *) working on boosting, a bunch of issues & PRs created. Right now I need > feedback on 'em. > *) Summing up my plans for HTM challenge, I'd like to do these > "projects"/papers - > -- ML benchmarks : > a repo with collection of usecases, datasets, algorithms and results. Ofc > we'd try HTM topping it! ;) > This should serve as an overview of practical HTM capabilities, comparison > for experts from other "AI schools of thought", and importantly source for > improvement ideas for nupic. > I plan to steal,cooperate with .research, .vision, .biodat, .nlp repos > etc.. > > -- HTM theory to other bio-inspired NN approaches comparison: > summing up the relation/differences on the interesting projects, ideas > that are around. I haven't found a complete, highlevel but broad overview, > which I can the pass to the bio/neuro-logists I try to cooperate with. > > -- Bio/Psycho-logical interdisciplinary example: > bringing HTM model to some bio/psych -logical condition > > Cheers, Mark > > >> 2. Working with the community in preparing and collecting goals / issues >> for the next great push for HTM.java. (Probably Serialization) >> 3. Working on a generalized JavaFX google mapping API (FXMaps) [1]. >> 4. Following #3, Working on Nostromo, an HTM Challenge entry [2]. >> 5. Integrating the Cognition Mission blog into my work routine. (Getting >> ready for next post) [3]. >> 6. Taking antacid pills to curb my anxiousness to hear Jeff Hawkins' next >> state of the HTM-union address & Fergal Byrne's next presentation on HTM >> technology [4]. >> >> [1] https://github.com/cogmission/FXMaps >> [2] https://github.com/cogmission/Nostromo >> [3] http://cogmission.ai >> [4] https://www.tums.com >> >> >> Peace, >> David >> > -- > Marek Otahal :o) > -- *With kind regards,* David Ray Java Solutions Architect *Cortical.io <http://cortical.io/>* Sponsor of: HTM.java <https://github.com/numenta/htm.java> [email protected] http://cortical.io
