集成学习

Integrated Learning

本文分析了当今机器学习现状,总结出三种学习体制:联结学习、遗传算法和符号学习。研究了三种体制各自的内涵和在智能活动中的地位及其相互间的联系。在此基础上,提出了综合不同学习方法的集成体制方案,并实现了一个基于解释的学习和基于相似学习的集成规则学习系统IR1。还讨论了联系学习与其它体制的局部集成。

The paper analyzes the state-of-the-art of machine learning research,and summarizes three learning paradigms:connectionist learning,genetic algorithm and symbolic approach.On the basis of this analysis,the author presents an integrated framework of making a cooperative use of different paradigms.The implementation IR1 is such an example of combining similarity-based learning and explanation-based learning.Also a possibility of integrating connectionist learning with other paradigms is explored here.