Re: [Pcsim-users] How to make an all-to-all projection

[Natschläger T. <Tho...@sc...>]

Hi:

Actually the RandomConnections with conn_prob = 1 is implemented such that no random generator is called (see file RandomConnections.cpp).
Hence RandomConnections( conn_prob = 1 ) seems a good way to implement a all-to-all connectivity.

Nevertheless it seems a good idea to include a AllToAllConections class in the PCSIM release.

Regards

  -thomas natschlager

-----Original Message-----
From: pcs...@li... on behalf of de...@ig...
Sent: Sat 9/13/2008 10:42 AM
To: pcs...@li...
Subject: Re: [Pcsim-users] How to make an all-to-all projection
 
Hi Mathieu,

Currently this is the only way to create all-to-all connectivity. Off  
course, it's not the optimal way since at each neuron pair a RNG is  
polled to decide whether a connection is to be formed. One can  
implement AllToAllConnectionIterator for this purpose (if it's not  
already there), or else
derive from ConnectionDecisionPredicate a TruePredicate which always  
decides True and then use it with:

proj = ConnectionsProjection( pop1, pop2, StaticSpikingSynapse(),
                               PredicateBasedConnections( TruePredicate() ) )


SimpleAllToAllWiringMethod is not meant for all-to-all connectivity  
projections.
WiringMethod classes are for optimization of connection creations in  
distributed mode for large networks, and are always used with  
ConnectionIterator. The actual logic is contained in the  
ConnectionIterator class.

regards,
Dejan




Quoting Mathieu Dubois <mat...@li...>:

> Hi everybody,
>
> This may seem a dumb question but I don't understand how can I make an
> all-to-all connection from one population to one other (the context is:
> connecting a reservoir to readout neurons).
>
> For the moment, I use a RandomConnections with a probability of 1. This is
> certainly not the best way. I have looked at Class
> SimpleAllToAllWiringMethod but I was unable to use it correctly.
>
> Could anyone provide a working example? I use a modest size reservoir so
> parallelization is not an issue.
>
> Thanks in advance (class AdvanceInfo),
> Mathieu
>
>
>
> -------------------------------------------------------------------------
> This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
> Build the coolest Linux based applications with Moblin SDK & win great prizes
> Grand prize is a trip for two to an Open Source event anywhere in the world
> http://moblin-contest.org/redirect.php?banner_id=100&url=/
> _______________________________________________
> Pcsim-users mailing list
> Pcs...@li...
> https://lists.sourceforge.net/lists/listinfo/pcsim-users
>



-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Pcsim-users mailing list
Pcs...@li...
https://lists.sourceforge.net/lists/listinfo/pcsim-users