Aleksander Klibisz, Grant Bruer, Catherine D. Schuman and James S. Plank

May, 2017

IJCNN: The International Joint Conference on Neural Networks

http://www.ijcnn.org/

View Article

Abstract

In recent years, research on neuromporphic computing platforms has focused on variable-structure, spiking network models. An important methodology for programming these networks is evolutionary optimization (EO), where thousands of networks are generated and then evaluated by determining fitness scores on specific tasks. Fitness scores guide the generation of new networks until a target fitness is achieved. One source of performance overhead during EO is the simulation of the task on each network to determine its fitness. To mitigate this source of overhead, we formulate the Static Fitness Prediction Task (SFPT), for predicting a network’s fitness without direct simulation. Our hypothesis is that we can use SFPT to predict network fitness sufficiently accurately to skip the simulation of a significant portion of networks during EO, thereby making the EO more efficient. We propose a data-driven approach to the SFPT on the neuromorphic model DANNA. Our approach transforms networks into directed graphs and extracts structural features to train an ancillary model for predicting the fitness of new networks. We analyze the extracted features and evaluate several predictive models to predict the fitness of networks for five tasks. Our results demonstrate a predictive capacity in these features and models. Our primary contribution is to demonstrate the utility of graph-level features extracted from variable-structure networks to predict network fitness and circumvent expensive simulations.

Citation Information

Text

author      A. Klibisz and G. Bruer and C. D. Schuman and J. S. Plank
title       Structure-based Fitness Prediction for the Variable-structure DANNA Neuromorphic Architecture
booktitle   IJCNN: The International Joint Conference on Neural Networks
month       May
year        2017
address     Anchorage

Bibtex

@INPROCEEDINGS{kbsp:17:sb,
    author = "A. Klibisz and G. Bruer and C. D. Schuman and J. S. Plank",
    title = "Structure-based Fitness Prediction for the Variable-structure DANNA Neuromorphic Architecture",
    booktitle = "IJCNN: The International Joint Conference on Neural Networks",
    month = "May",
    year = "2017",
    address = "Anchorage"
}