NATURAL LANGUAGE UNDERSTANDING;
NON-SYNONYMIC PARAPHRASING;
SEMANTIC RELEVANCE;
COMPONENTS OF MEANING;
D O I:
暂无
中图分类号:
TP18 [人工智能理论];
学科分类号:
081104 ;
0812 ;
0835 ;
1405 ;
摘要:
A large set of theories has been proposed in recent years, covering one or more aspects of natural language understanding. Few of them could have been integrated successfully into one working system. SAPFO is a paraphrasing and question-answering system, designed to experiment with language as a whole. Based on a psycholinguistic experiment with re-narration of stories, an experimental system for paraphrasing of Slovak has been built. The paraphraser serves as a computer tool for linguistic analysis. It enables to test particular linguistic hypotheses or concepts, and to watch their behavior in mutual interaction with other subsystems of language and on large corpora. We claim that more attention should be paid to non-synonymical relationships in language. An algebra of so called synsemantic transformations enables to match semantically relevant constructions of quite different surface forms.