article · 1992
Learning with Limited Numerical Precision Using the Cascade-Correlation Algorithm
Markus Hoehfeld Scott, Scott E. Fahlman
Cited by 0Abstract and details only
Abstract
A key question in the design of specialized hardware for simulation of neural networks is whether fixedpoint arithmetic of limited numerical precision can be used with existing learning algorithms. We present an empirical study of the effects of limited precision in Cascade-Correlation networks on three different learning problems. We show that learning can fail abruptly as the precision of network weights or weight-update calculations is reduced below 12 bits. We introduce techniques for dynamic rescaling and probabilistic rounding that allow reliable convergence down to 6 bits of precision, with only a gradual reduction in the quality of the solutions. This research was supported by Siemens Corporation. The experiments described here were conducted during a visit by Markus Hoehfeld to Carnegie Mellon in the fall of 1990. Markus Hoehfeld's permanent address is Siemens AG, ZFE IS INF 2, Otto-Hahn-Ring 6, W-8000 Munich 83, Germany. Keywords: Artificial Intelligence, Learning, Neural n...
Cite this paper
Scott, M. H., & Fahlman, S. E. (1992). Learning with Limited Numerical Precision Using the Cascade-Correlation Algorithm.
Scott, Markus Hoehfeld, and Scott E. Fahlman. “Learning with Limited Numerical Precision Using the Cascade-Correlation Algorithm.” 1992.
Scott, Markus Hoehfeld, and Scott E. Fahlman. 1992. “Learning with Limited Numerical Precision Using the Cascade-Correlation Algorithm.”
Scott, M.H. and Fahlman, S.E. (1992) ‘Learning with Limited Numerical Precision Using the Cascade-Correlation Algorithm’.
@article{scott1992learning,
title = {{Learning with Limited Numerical Precision Using the Cascade-Correlation Algorithm}},
author = {Scott, Markus Hoehfeld and Fahlman, Scott E.},
year = {1992}
}Made from the index record. Check names, capitalisation and page numbers against the paper. Need another style? AI for Research has over 10,000.