|
|||||
|
|
| 从规划解中学习一阶派生谓词规则 | |
| 引用本文： | 饶东宁,蒋志华,姜云飞,刘强.从规划解中学习一阶派生谓词规则[J].计算机学报,2010,33(2). |
| 作者姓名： | 饶东宁 蒋志华 姜云飞 刘强 |
| 作者单位： | 1. 广东工业大学计算机学院,广州,510090;中山大学信息科技学院软件研究所,广州,510275 2. 暨南大学计算机科学系,广州,510632 3. 中山大学信息科技学院软件研究所,广州,510275 |
| 基金项目： | 国家自然科学基金(60173039);;广东工业大学博士启动基金项目(093032)资助~~ |
| 摘 要： | 派生谓词是描述动作非直接效果的主要方式.但是由人类专家设计的派生谓词规则(即领域理论)不能保证总是正确或者完备的,因此有时很难解释一个观察到的规划解为什么是有效的.结合归纳学习与分析学习的优点,文中提出一种称为FODRL(First-Order Derived Rules Learning)的算法,在不完美的初始领域理论的引导下从观察到的规划解中学习一阶派生谓词规则.FODRL基于归纳学习算法FOIL(First-Order Inductive Learning),最主要的改进是可以使用派生谓词的激活集来扩大搜索步,从而提高学习到的规则的精确度.学习过程分为两个步骤:先从规划解中提取训练例,然后学习能够最好拟合训练例和初始领域理论的一阶规则集.在PSR和PROME-LA两个派生规划领域进行实验,结果表明,在大部分情况下FODRL比FOIL(甚至包括其变型算法FOCL)学习到的规则的精确度都要高. |
| 关 键 词： | 人工智能 智能规划 派生谓词规则 归纳学习 激活集 |
Learning First-Order Rules for Derived Predicates from Plan Examples |
|
| RAO Dong-Ning,JIANG Zhi-Hua,JIANG Yun-Fei,LIU Qiang.Learning First-Order Rules for Derived Predicates from Plan Examples[J].Chinese Journal of Computers,2010,33(2). | |
| Authors: | RAO Dong-Ning JIANG Zhi-Hua JIANG Yun-Fei LIU Qiang |
| Affiliation: | Faculty of Computer/a>;Guangdong University of Technology/a>;Guangzhou 510090;Software Research Institute/a>;School of Information Science and Technology/a>;Sun Yat-Sen University/a>;Guangzhou 510275;Department of Computer Science and Technology/a>;Jinan University/a>;Guangzhou 510632 |
| Abstract: | Derived predicates are a natural way to depict indirect effects of domain actions,and their truth values in the current state are inferred from that of other predicates via domain rules.However,domain rules designed by human experts cannot be guaranteed to be correct or complete.So it is often difficult to explain why an observed plan is valid under imperfect domain rules.Combining inductive learning with analytical learning,in this paper,we develop an algorithm called FODRL(First-Order Derived Rules Learni... |
| Keywords: | artificial intelligence automated planning derived predicate rules inductive learning activation set |
| 本文献已被 CNKI 万方数据 等数据库收录！ | |