Titre :  [FORTHCOMING] Analogia and Intersector - games as a basis for automatic analogy analysis 

Auteurs : Mathieu Lafourcade
, Nathalie Le Brun, Jérémie Roux,  

Revue : Technology and Innovation 

Numéro : Forthcoming papers 

Volume :  

Date : 2024/05/31 

DOI : TBA 

ISSN : 2399-8571 

Résumé : computer system’s ability to analyze texts semantically involves interpreting figurative content. Analogy is often used to convey new ideas through similarity with known ideas. Comparison and metaphor can be modeled as an analogical square A : B :: C : D (meaning A is to B what C is to D) with one or more variables whose most relevant values must be found. This article presents two games for collecting lexical data related to analogies. The first game, Analogia, is an adaptation of JeuxDeMots to analogies, where the player has to provide the most relevant answers for an analogy to a variable. For example, find x for “coal is to black what snow is to x”. The second game, Intersector, involves answering questions to find out what the 4 terms in an analogy have in common. The terms proposed are those taken from the analogies in the Analogia game. All the data collected by these games makes it possible to automatically resolve analogies, and thus interpret figurative content. 

Éditeur : ISTE OpenScience