Even simple neural nets cannot be trained reliably with a polynomial number of examples

International 1989 Joint Conference on Neural Networks Pub Date : 1989-12-01 DOI:10.1109/IJCNN.1989.118691

H. Shvaytser

引用次数: 7

Abstract

A variation of L.G. Valiant's 'PAC' model of learnability (Commun. ACM, vol.27, no.11, p.1134-42, 1984; Proc. 9th Int. Joint Conf. Artif. Intell., Aug. 1985) is used to investigate the learning power of artificial neural nets with threshold nodes. It is shown that there are cases where simple nets require an exponential number of training examples for reliably determining their sets of parameters. Polynomially many training examples may not be enough to determine the set of parameters even for a net of three threshold nodes, if it has to perform reliably in two different environments.<>

查看原文本刊更多论文

即使是简单的神经网络也不能用多项式数量的样本进行可靠的训练

这是L.G. Valiant的“PAC”易学性模型的变体。美国计算机学会，第27卷，第27期。11，第1134-42页，1984;第9段联合会议智能。(1985年8月)用于研究具有阈值节点的人工神经网络的学习能力。结果表明，在某些情况下，简单网络需要指数数量的训练样例才能可靠地确定其参数集。如果它必须在两个不同的环境中可靠地执行，那么即使对于三个阈值节点的网络，多项式地许多训练样例也可能不足以确定参数集。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

International 1989 Joint Conference on Neural Networks

自引率

0.00%

发文量