Edward, Did you read Izhikevich's papers (specifically, [1])? They explore the model of polychronization, where cell assemblies are formed in different ways depending on temporal shifts of firings of neurons that initiate assembly's formation. He has some experimental estimations, but they are drawn on untrained (content-wise) networks, so it's unclear if they are useful (in paper I referenced he gives a figure of about 5000 groups in a network of 1000 neurons with 100 connections per neuron and delays of 1 to 20 synch spans). Also definition of polychronous group is too wide and seem to include stable patterns of behavior of any complexity (or simplicity...), which doesn't correspond well to cell assemblies. For example, paper [2] gives a figure of 3000 groups on a network of 200 neurons.
Still I don't see how it could help with induction: as I wrote in parallel thread, number of connections is a strong limitation that can't be overcome with dynamics without some kind of global identifier coding. _______________ [1] Izhikevich E.M. (2006). Polychronization: Computation With Spikes. **<http://vesicle.nsi.edu/users/izhikevich/publications/spnet.htm> [2] Helene Paugam-Moisy, Regis Martinez and Samy Bengio. (2007). A supervised learning approach based on STDP and polychronization in spiking neuron networks. On 10/22/07, Edward W. Porter <[EMAIL PROTECTED]> wrote: > > Vladmir, > > I think a very important issue, ist the one about how much you can > multiplex the number of cell assemblies a neuron is in. If X is the total > number of neurons, and M is the number of neurons in a cell assembly, as in > one of your earlier posts, and you assume even distribution of > connections, the total number of cell assemblies would be X/M. But with > multiplexing it might be substantially higher. I have read or heard > multiple people say the number can be much higher, but I have heard no one > put a number or mathematical formula on exactly how much higher. Do you > know anything about that? > > Edward W. Porter > Porter & Associates > 24 String Bridge S12 > Exeter, NH 03833 > (617) 494-1722 > Fax (617) 494-1822 > [EMAIL PROTECTED] > > -----Original Message----- > *From:* Vladimir Nesov [mailto:[EMAIL PROTECTED] > *Sent:* Sunday, October 21, 2007 6:51 PM > *To:* agi@v2.listbox.com > *Subject:* Re: [agi] Human memory and number of synapses.. P.S. > > On 10/21/07, Edward W. Porter <[EMAIL PROTECTED]> wrote: > > > > Vladimir, > > > > > > > > Yes, if a concept is defined by its associations, and if a significant > > subset of them somewhat distinguish a concept, it would seem only natural > > that links between associations of nodes A and node could help the two > > concepts find each other in a large, high dimensional space. > > > > Or, specifically, simple notion of neuron-level interaction doesn't seem > to provide support for induction... > > > -- > Vladimir Nesov mailto: [EMAIL PROTECTED] > ------------------------------ > This list is sponsored by AGIRI: http://www.agiri.org/email > To unsubscribe or change your options, please go to: > http://v2.listbox.com/member/?&; > > ------------------------------ > This list is sponsored by AGIRI: http://www.agiri.org/email > To unsubscribe or change your options, please go to: > http://v2.listbox.com/member/?&; > -- Vladimir Nesov mailto:[EMAIL PROTECTED] ----- This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=8660244&id_secret=56150010-933233
