Abstract | ||
---|---|---|
Referring expressions aim to distinguish an object or a set of objects within a collection in terms of their properties. Within the Natural Language Generation field, Referring Expression Generation is one of the key problems, where graduality plays a very relevant role. In this paper we focus on quantified relational properties of objects, which are defined through their relations to other objects, and are of special interest in the generation of referring expressions. Our proposal involves fuzzy sets in the definition of relations, properties, and generalized quantification for such purpose. General quantified relation patterns are introduced, of which particular examples are To be related to at least n objects of type A, To be related to more objects of type A than objects of type B, To be related to all objects of type A except n, and To be next to many more objects of type A than objects of type B, among many others. |
Year | DOI | Venue |
---|---|---|
2022 | 10.1109/FUZZ-IEEE55066.2022.9882695 | 2022 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE) |
Keywords | DocType | ISSN |
Quantified referring expressions,fuzzy generalized quantification,quantified relational properties,graduality | Conference | 1544-5615 |
ISBN | Citations | PageRank |
978-1-6654-6711-7 | 0 | 0.34 |
References | Authors | |
15 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Nicolás Marín | 1 | 0 | 0.34 |
Gustavo Rivas-Gervilla | 2 | 1 | 2.39 |
Daniel Sánchez | 3 | 967 | 60.29 |