A CLP-Based, Diagnosticity-Driven System for Concept Combinations

Abstract

Diagnosticity operates as an important selection criterion for several computational models of concept combination. Unfortunately, it has not been clear how the diagnosticity of property and relational predicates of the concepts combined can be formalized and quantified. Using an information retrieval method we compute, in a uniform manner, diagnosticity values of concepts predicates. We go on to present a reasoning system that attempts to create meaningful interpretations of novel nounnoun combinations. The system is based solely on diagnostic predicates values and a set of constraint satisfaction rules. We show the effectiveness and plausibility of our methods and discuss their potential. 1

Cite

Text

Tagalakis et al. "A CLP-Based, Diagnosticity-Driven System for Concept Combinations." International Joint Conference on Artificial Intelligence, 2005.

Markdown

[Tagalakis et al. "A CLP-Based, Diagnosticity-Driven System for Concept Combinations." International Joint Conference on Artificial Intelligence, 2005.](https://mlanthology.org/ijcai/2005/tagalakis2005ijcai-clp/)

BibTeX

@inproceedings{tagalakis2005ijcai-clp,
  title     = {{A CLP-Based, Diagnosticity-Driven System for Concept Combinations}},
  author    = {Tagalakis, Georgios and Ferrari, Daniela and Keane, Mark T.},
  booktitle = {International Joint Conference on Artificial Intelligence},
  year      = {2005},
  pages     = {1585-1586},
  url       = {https://mlanthology.org/ijcai/2005/tagalakis2005ijcai-clp/}
}