>>2. Will the new TM implementation be encapsulated in more "biological-like" structures? I noticed a difference between the work Marek has done and the current CLA implementation in terms >>of the discrete biological modeling?
>I'm not sure I fully understand your question. It will be closer to the white paper, without the non-biological additions that currently exist in the TP implementation. Also, it will be a little easier to map to the >biology than the white paper description (and will incorporate a couple minor changes to make it more biologically feasible). But at the end of the day, the white paper description draws almost directly from >our understanding of the biology. My comments could have be due to the difference in coding styles when coding in Java vs. Python? What I was saying was I didn't notice an overt code files/classes named Synapse; Dendrite; DistalDendrite; Layer etc. (within the "research" directory where TP.py and spacial_pooler.py exist) for which there are explicit references in the white paper? David On Tue, Jul 15, 2014 at 12:36 PM, Chetan Surpur <[email protected]> wrote: > Hi David, > > 2. Will the new TM implementation be encapsulated in more > "biological-like" structures? I noticed a difference between the work Marek > has done and the current CLA implementation in terms of the discrete > biological modeling? > > > I'm not sure I fully understand your question. It will be closer to the > white paper, without the non-biological additions that currently exist in > the TP implementation. Also, it will be a little easier to map to the > biology than the white paper description (and will incorporate a couple > minor changes to make it more biologically feasible). But at the end of the > day, the white paper description draws almost directly from our > understanding of the biology. > > - Chetan > > On July 14, 2014 at 9:58:28 PM, cogmission1 . ([email protected]) > wrote: > > Hi Matthew (Lohbihler), Yuwei, Chetan, > > I apologize for my lack of a prompt response. Thank you all for your > responses! > > @Matthew - I initially wrote about my interest in porting the existing > implementation to Java in another thread in which Matthew (Taylor) > suggested I have a look at Marek Otahal's project on github (listed below). > > >> If you're interested in porting to Java, you might have a look at > Marek Otahal's "htm-cla" project on Github [1]. Marek is one of our main > community contributors, and before we open-sourced NuPIC, he wrote up his > implementation of the CLA White Paper algorithms in Java (that's how I > found him!). > > >> [1] https://github.com/breznak/htm-cla > > If you're interested in the Java port, you might have a look at this too. > I'm still building my understanding of the SP/TM/(and future TP) > mechanisms, and would like to wait for the new TM mechanism that Chetan is > developing and as well as the new TP implementation. I think I would like > to understand the various mechanisms very thoroughly before "diving in" to > code - but I'm still available for discussion in the meantime; we should > keep in touch and thank you for your offer! > > In addition, I'm still reviewing the implementation of Marek's which will > also help build an understanding. I will probably develop an additional > implementation, though not as a comment or critique on Marek's efforts, > but just so that I can "shore" up my own understanding and go beyond the > core mechanisms into the encoders, classifiers, and other peripheral > support modules with the Java port. > > @Yuwei, Chetan - thank you for your clarifications and sharing the > progress on your current projects. I wanted to ask another question. > > 1. Outside of the CLA algorithm implementation there are the swarming > mechanism, the encoders and classifiers. Are these considered to be the > difference between the "core" algorithm and the Online Prediction Framework > (OPF)? > > 2. Will the new TM implementation be encapsulated in more > "biological-like" structures? I noticed a difference between the work Marek > has done and the current CLA implementation in terms of the discrete > biological modeling? > > Cheers, > David > > > On Mon, Jul 14, 2014 at 12:43 PM, Chetan Surpur <[email protected]> > wrote: > >> David, >> >> I'd also like to mention that I'm beginning work on a simplified >> temporal memory (what you called sequence memory) implementation, that's >> closer to the white paper description. This work should be completed in the >> near future. However, you can still definitely port the existing temporal >> memory (currently called TP) implementation to Java if you don't want to >> wait for the simplified one. >> >> - Chetan >> >> >> On July 14, 2014 at 9:02:58 AM, Yuwei Cui ([email protected]) wrote: >> >> Hi David, >> >> Each layer will be implementing sequence memory (we now call it temporal >> memory). The goal is to predict future input based on past inputs. Spatial >> pooling and sequence memory will not be combined. If you have read the >> white paper, the spatial pooling occurs on the proximal dendrite, which can >> drive cells into "active states", and sequence memory relies on inputs on >> distal dendrite, which will turn cells into "predicted states" >> >> Our current working hypothesis is temporal pooling occurs both between >> layer 4 and layer 3, and across regions. The outcome of temporal pooling is >> more stable representation higher in the hierarchy. There will be an >> emphasis on the stability soon, we already have a working temporal pooling >> algorithm and will release it in the near future. The inter-regional >> feedback seems a little further down the road. >> >> Yuwei >> >> >> On Sun, Jul 13, 2014 at 3:12 PM, cogmission1 . < >> [email protected]> wrote: >> >>> Hey Matt, >>> >>> I just watched the "Fireside chat" about Jeff's next TP ideas, and not >>> to be argumentative (just looking to be as clear as I can be while watching >>> things happen from the outside), but it appears as if he's going to >>> introduce sequence memory, and temporal pooling between each layer? I think >>> definitely between within layers 3 and 4? Also, it appears as if the SP and >>> the Sequence Memory will be combined? Additionally, it looks like there >>> will be some "inter-regional" feedback/predictions introduced between a >>> subsequent region's layer 6 back into layer 5 of the previous layer. It >>> maybe looks like some drastic overall changes may be occurring - maybe due >>> to an effort toward emphasizing concepts of stability and inter-regional >>> feedback?. Any news as to when we could see these changes rolled out - I >>> can hardly wait...? :-P >>> >>> >>> >>> >>> On Sat, Jul 12, 2014 at 12:08 PM, Matthew Taylor <[email protected]> >>> wrote: >>> >>>> You can definitely dive in now. The new temporal memory work isn't >>>> going to disrupt the current architecture at all, as far as I know. >>>> >>>> Start here: https://github.com/numenta/nupic/wiki/Learning-NuPIC >>>> >>>> --------- >>>> Matt Taylor >>>> OS Community Flag-Bearer >>>> Numenta >>>> >>>> >>>> On Sat, Jul 12, 2014 at 9:42 AM, cogmission1 . < >>>> [email protected]> wrote: >>>> >>>>> Hi Everyone, >>>>> >>>>> I'm at the point now where I would begin to delve into the NuPIC code >>>>> more deeply (I want to port it to Java as a way to exercise my >>>>> understanding too), but I was wondering if there is any point to it prior >>>>> to the new TP/Sequence Memory implementation? Should I maybe wait? >>>>> >>>>> Thanks >>>>> >>>>> David >>>>> >>>>> _______________________________________________ >>>>> 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 >>> >>> >> >> >> -- >> -- >> Yuwei Cui >> >> Algorithm Internship, Numenta Inc. >> >> PhD Candidate, Neuroscience and Cognitive Science >> >> University of Maryland, College Park, MD, 20742 >> >> Homepage: http://terpconnect.umd.edu/~ywcui/ >> _______________________________________________ >> 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 > > > _______________________________________________ > 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
