【24h】

Commonsense Knowledge Acquisition from WordNet

机译:来自Wordnet的型号知识获取

获取原文

摘要

This paper proposes a semiautomatic method for generating commonsense axioms. The method starts with a set of commonsense rules as seed input, then it applies a metarule to these commonsense rules to generate commonsense axioms by searching over the semantically enhanced version of WordNet. The WordNet glosses have been syntactically and semantically parsed and transformed into semantic triples. The algorithm searches the WordNet for all concepts that have a property specified in a seed rule thus instantiating the seed rule. Axioms are generated linking these concepts with the commonsense rule by using semantic triples. The results show that in one particular instance the algorithm generated 1051 axioms by using a seed of 35 commonsense rules with approximately 98% accuracy.
机译:本文提出了一种用于产生勤义公理的半自动方法。该方法从一组致辞规则开始作为种子输入,然后它将元素应用于这些偶数规则,以通过搜索TOWNNET的语义增强版本来生成致辞原理。 Wordnet光泽已在句子上和语义上解析并转换为语义三元。该算法搜索WordNet的所有概念,其中包含在种子规则中指定的属性,从而实例化种子规则。通过使用语义三元组将这些概念与致辞规则联系起来的公理。结果表明,在一个特定的实例中,算法通过使用35个致辞规则的种子来生成1051个公理,精度大约98%。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号