JuNe - The Juelich Network

From KIP Wiki
Revision as of 13:52, 19 April 2012 by Schrader (talk | contribs) (Created page with "=== Overview === Sensory cortices undergo a massive bombardment of input. Network dynamics upon thalamocortical stimulation shows a stereotyped spread of activation which can ...")
(diff) ↠Older revision | Latest revision (diff) | Newer revision → (diff)
⧼kip-jumptonavigation⧽⧼kip-jumptosearch⧽

Overview

Sensory cortices undergo a massive bombardment of input. Network dynamics upon thalamocortical stimulation shows a stereotyped spread of activation which can be largely explained by the connection probabilities. The interplay between stimulation, feed-forward and feedback, as well as recurrent connections causes asynchronous irregular activity and oscillations. This model aims at a massively parallel simulation of interconnected local cortical networks, each comprising in the order of 80,000 neurons. Ideally, the model should be operational on the wafer-scale system as well as on the Jugene cluster.

A description of this project can also be found here: https://brainscales.kip.uni-heidelberg.de/internal/wiki/index.php/WP_5_index

Developing institutions

FZJ

Developers

Current developers are S. van Albada, M. Diesmann and S. Schrader

Results of the ESS workshop, Oct 2011

The model was presented at the Using the ESS + Neuromorphic hardware workshop in Dresden, Oct 4th, 2011 by Sacha van Albada. The slides of the presentation can be found here

  • An initial model of JuNe has been implemented in PyNN and is operational on the ESS.
  • The original neuron model has been transferred to the hardware neuron (AdEx).
  • The necessity of massive Poisson input has to be discussed in the presence of hardware limitations. Several strategies to ease bandwidth have been proposed and implemented in PyNN.
  • Synaptic homogeneities (delays, weights) may surpass hardware limitations, but the former can possibly be approximated with distributed time constants, if necessary.