Optimal Connectivity In Hardware-Targeted MLP Networks

Alexander Rast, SR Welbourne, X. Jin, Stephen Furber

Research output: Chapter in Book/Conference proceedingConference contributionpeer-review

Abstract

n large neural networks, partial connectivity is both biologically plausible and a matter of necessity when targetting a hardware implementation. We are using the SpiNNaker neural chip multiprocessor to model such networks as a drop-in replacement for the Lens network simulator. For the popular MLP network, a theoretical model of the relation between connectivity, network size and gain in the activation function provides a method to set these parameters to near-optimal values. Using the model, we run a series of network simulations in Lens, permuting the parameters to explore the effects in 2 networks of different size and application. Initial test results show a clear connectivity-gain relation and a benefit to partial connectivity in both networks, with optimal hidden-output connectivity values ranging from ~10% - ~30% depending on the network type. We show that optimal connectivity-gain settings reduce training time, minimising error oscillations during learning. Preliminary analysis also suggests that while very low connectivities may improve error they may also result in decreased adaptivity to new inputs or component failure. These results in combination with the theoretical relation give a method for determining reasonable initial connectivity and gain values \textit{at design time} for an MLP network, allowing more efficient use of hardware resources such as SpiNNaker and faster simulations in any software environment. They also suggest a different way of considering the problem of MLP network design: rather than specify a fixed number of neurons, specify a fixed number of connections and vary the number of neurons to reach optimal connectivity.
Original languageEnglish
Title of host publicationProceedings 2009 International Joint Conference on Neural Networks, IJCNN2009
Pages2619-2626
Number of pages8
Publication statusPublished - Jun 2009
EventInternational Joint Conference on Neural Networks, IJCNN2009 - Atlanta
Duration: 14 Jun 200919 Jul 2010

Conference

ConferenceInternational Joint Conference on Neural Networks, IJCNN2009
CityAtlanta
Period14/06/0919/07/10

Fingerprint

Dive into the research topics of 'Optimal Connectivity In Hardware-Targeted MLP Networks'. Together they form a unique fingerprint.

Cite this