摘要
本文介绍一个具有自学习功能的软件自动化系统NDSAiL。该系统能自动从示例中学习问题分解方法,并用于自动生成部分通用算法.该系统还能自动学习基本算法及算法优化方法.在算法构架学习背景下,有效地解决了解释学习的操作性准则问题;并针对一类问题解决了归纳学习结果的正确性问题以及unfold/fold转换技术的完全自动化问题.
A software automation system with automatic learning mechanisms is described, which can learn problem decomposion methods from training examples automatically,and use them in generating general algorithms. It can also learn, fundamental algorithms and optimizing methods. In algorithm scheme learning, the problem of operationality criterion is solved successfully; the correctness of inductive learning and full automation of unfold/fold transformation are solved for a limited class of probelms
出处
《计算机学报》
EI
CSCD
北大核心
1992年第11期819-829,共11页
Chinese Journal of Computers
基金
国家863高技术项目基金
关键词
机器学习
软件自动化
设计
Machine learning, EBL, software automation, algorithm synthesis.
