Conceptual structures: information processing in mind and machine
Conceptual structures: information processing in mind and machine
Tools and methods for computational lexicology
Computational Linguistics - Special issue of the lexicon
Semantic distance in conceptual graphs
Conceptual structures
Determining similarity and inferring relations in a lexical knowledge base
Determining similarity and inferring relations in a lexical knowledge base
From a children's first dictionary to a lexical knowledge base of conceptual graphs
From a children's first dictionary to a lexical knowledge base of conceptual graphs
Redundancy: helping semantic disambiguation
COLING '98 Proceedings of the 17th international conference on Computational linguistics - Volume 1
Semantically significant patterns in dictionary definitions
ACL '86 Proceedings of the 24th annual meeting on Association for Computational Linguistics
Extracting semantic hierarchies from a large on-line dictionary
ACL '85 Proceedings of the 23rd annual meeting on Association for Computational Linguistics
Extraction of semantic information from an ordinary English dictionary and its evaluation
COLING '88 Proceedings of the 12th conference on Computational linguistics - Volume 2
Structural patterns vs. string patterns for extracting semantic information from dictionaries
COLING '92 Proceedings of the 14th conference on Computational linguistics - Volume 2
Using information content to evaluate semantic similarity in a taxonomy
IJCAI'95 Proceedings of the 14th international joint conference on Artificial intelligence - Volume 1
Non-classical lexical semantic relations
CLS '04 Proceedings of the HLT-NAACL Workshop on Computational Lexical Semantics
Hi-index | 0.00 |
Type hierarchies are important structures used in knowledge stores to enumerate and classify the entities from a domain of interest. The hierarchical structure establishes de facto a "similarity space" in which the elements of a same class are considered close semantically, as they share the properties of their superclass. An important task in Natural Language Processing (NLP) is sentence understanding. This task relies partly on comparing the words in the sentence among each other as well as to the words in previous sentences and to words in a knowledge store. A type hierarchy consisting of words and/or word senses can be useful to facilitate these comparisons and establish which words are semantically related. The problems of using a type hierarchy for evaluating semantic distance come from its dependency on the available words of a specific language, and on the arbitrariness of its classes and of its depth, which leads to the development of semantic distance measures giving arbitrary results. We propose a way to extend the type hierarchy, to give more flexibility to the "similarity space", by including nonlexical concepts defined around relations other than taxonomic ones. We also suggest a method for discovering these non-lexical concepts in texts, and present some results.