Pasar al contenido principal

SERVICIOS INTELIGENTES COORDINADOS PARA AREAS INTELIGENTES ADAPTATIVAS

PID2021-123673OB-C32

•
Nombre agencia financiadora Agencia Estatal de Investigación
Acrónimo agencia financiadora AEI
Programa Programa Estatal para Impulsar la Investigación Científico-Técnica y su Transferencia
Subprograma Subprograma Estatal de Generación de Conocimiento
Convocatoria Proyectos de I+D+i (Generación de Conocimiento y Retos Investigación)
Año convocatoria 2021
Plan Estatal 2021-2023 Plan Estatal de Investigación Científica y Técnica y de Innovación
Centro beneficiario UNIVERSIDAD REY JUAN CARLOS

Resultados relacionados

Resultados totales (Incluyendo duplicados): 2
Encontrada(s) 1 página(s)

DIGITAL.CSIC. Repositorio Institucional del CSIC
Publicación Otros (Other)

Opportunities and Challenges of the Use of Forgetting in Symbolic XAI

  • Arias, Joaquín
  • Degli Esposti, Sara
  • Hernández-Pérez, Jose Walter
  • Fidilio-Allende, Luciana
  • Ossowski, Sascha
The increasing use of artificial intelligence systems (especially those based on black-box algorithms) in decision models that affect humans is leading to an increase in demand for explainable AI (XAI) and models aligned with human values and norms. Recently, symbolic reasoning approaches, such as Answer Set Programming (ASP), have re-emerged as an alternative solution to create transparent and interpretable models more easily auditable. However, this explainability may expose private information (with ethical and legal implications). Although it is possible to manipulate the justifications (to hide sensitive information), ASP forgetting techniques are gaining relevance because they allow to remove such information from the models as well, thus allowing to audit them without compromising the users’ privacy and comply with GDPR data minimization principle. On the other hand, black box models cannot provide such explanations, so to ensure privacy preservation a naive approach is to learn Machine Learning (ML) models using a dataset from which sensitive attributes are removed, but this solution may affect accuracy. In this paper we present Privacy-ML, a complex scheme that is trained with sensitive information but then does not require such information when performing classification (improving the accuracy of naive approaches). Additionally, we have identified that a Privacy-ML based on Inductive Logic Programming can also be used by a malicious agent in combination with forgetting to obfuscate their deceptive profiling practices in such a way that it cannot be directly detected by someone reading the explanations it produces or by auditing the model., This work has been supported by grant VAE: TED2021-131295B-C33/C31 funded by MCIN/AEI/ 10.13039/501100011033 and by the “European Union NextGenerationEU/PRTR“, by grant COSASS: PID2021-123673OB-C32 funded by MCIN/AEI/ 10.13039/501100011033 and by “ERDF A way of making Europe”., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publicación Otros (Other)

Learning the Value Systems of Societies from Preferences

  • Holgado-Sánchez, Andrés
  • Holger Billhardt
  • Ossowski, Sascha
  • Degli Esposti, Sara
Aligning AI systems with human values and the value-based preferences of various stakeholders (their value systems) is key in ethical AI. In value-aware AI systems, decision-making draws upon explicit computational representations of individual values (groundings) and their aggregation into value systems. As these are notoriously difficult to elicit and calibrate manually, value learning approaches aim to automatically derive computational models of an agent’s values and value system from demonstrations of human behaviour. Nonetheless, social science and humanities literature suggest that it is more adequate to conceive the value system of a society as a set of value systems of different groups, rather than as the simple aggregation of individual value systems. Accordingly, here we formalize the problem of learning the value systems of societies and propose a method to address it based on heuristic deep clustering. The method learns socially shared value groundings and a set of diverse value systems representing a given society by observing qualitative value-based preferences from a sample of agents. We evaluate the proposal in a use case with real data about travelling decisions., This work is supported by grant VAE: TED2021-131295B-C33 funded by MCIN/AEI/10.13039/501100011033 and by “European Union NextGenerationEU/PRTR”, by grant COSASS: PID2021-123673OB-C32 funded by MCIN/AEI/10.13039/501100011033 and by “ERDF A way of making Europe”, and by AGROBOTS Project of University Rey Juan Carlos funded by Community of Madrid, Spain., Peer reviewed