Autor:
Fidilio-Allende, Luciana

Foto de perfil

E-mails conocidos

luciana.fidilio@urjc.es

Fecha de nacimiento

Proyectos de investigación

Unidades organizativas

Puesto de trabajo

Apellidos

Fidilio-Allende

Nombre de pila

Luciana

Nombre

Nombres alternativos

Afiliaciones conocidas

Universidad Rey Juan Carlos, Spain

Páginas web conocidas

http://www.ia.urjc.es/gia/luciana-fidilio/
Página completa del ítem
Notificar un error en este autor

Filtros

20241
Restablecer filtros

Ajustes

Resultados de la búsqueda

Mostrando 1 - 1 de 1
  • Miniatura
    Artículo
    fCASP: A forgetting technique for XAI based on goal-directed constraint ASP models
    Fidilio-Allende, Luciana; Arias, Joaquín. Actas de las XXIII Jornadas de Programación y Lenguajes (PROLE 2024), 2024-06-17.
    Artificial Intelligence systems based on machine learning are increasingly used to make decisions that directly affect humans, but they are not able to explain those decisions. On the other hand, Artificial Intelligence systems based on Constrained Answer Set Programming (CASP) provide human-readable justifications and their models can be audited and/or adapted, e.g., to ensure that they are value-aware. While this explainability is a legal (and ethical) requirement, it can lead to a leak of sensitive information, for example in cases of victims of gender-based violence. Although explanations can be manipulated to avoid leaks, when adapting the models, the application of techniques such as forgetting is required. However, current forgetting techniques are mostly only applied in propositional ASP programs, and they have limitations dealing with even loops. In this paper, we present preliminary results of a new forgetting technique, called fCASP , which can be successfully applied to examples that existing techniques are not able to solve correctly. fCASP is based on the dual rules of s(CASP), a goal-directed CASP reasoner, and therefore, we believe that it can be applied to generic CASP programs without grounding. We have validated our proposal by solving flagship examples from the literature, and we plan to use this technique in the context of school places allocation while preserving the privacy of victims of gender-based violence.