Current research in Explainable AI includes post-hoc explanation methods that focus on building transparent explaining agents able to emulate opaque ones. Such agents are naturally required to be accurate and trustworthy. However, what it means for an explaining agent to be accurate and trustworthy is far from being clear. We characterize accuracy and trustworthiness as measures of the distance between the formal properties of a given opaque system and those of its transparent explanantes. To this aim, we extend Probabilistic Computation Tree Logic with operators to specify degrees of accuracy and trustworthiness of explaining agents. We also provide a semantics for this logic, based on a multi-agent structure and relative model-checking algorithms. The paper concludes with a simple example of a possible application.
Modelling Accuracy and Trustworthiness of Explaining Agents / A. Termine, G. Primiero, F.A. D'Asaro (LECTURE NOTES IN ARTIFICIAL INTELLIGENCE). - In: Logic, Rationality, and Interaction / [a cura di] S. Ghosh, T. Icard. - [s.l] : Springer Science and Business Media, 2021. - ISBN 978-3-030-88707-0. - pp. 232-245 (( Intervento presentato al 8. convegno International Workshop on Logic, Rationality and Interaction tenutosi a Xi'An nel 2021 [10.1007/978-3-030-88708-7_19].
Modelling Accuracy and Trustworthiness of Explaining Agents
A. TerminePrimo
;G. PrimieroSecondo
;F.A. D'AsaroUltimo
2021
Abstract
Current research in Explainable AI includes post-hoc explanation methods that focus on building transparent explaining agents able to emulate opaque ones. Such agents are naturally required to be accurate and trustworthy. However, what it means for an explaining agent to be accurate and trustworthy is far from being clear. We characterize accuracy and trustworthiness as measures of the distance between the formal properties of a given opaque system and those of its transparent explanantes. To this aim, we extend Probabilistic Computation Tree Logic with operators to specify degrees of accuracy and trustworthiness of explaining agents. We also provide a semantics for this logic, based on a multi-agent structure and relative model-checking algorithms. The paper concludes with a simple example of a possible application.File | Dimensione | Formato | |
---|---|---|---|
Termine,Primiero,Dasaro.pdf
Open Access dal 05/10/2022
Tipologia:
Post-print, accepted manuscript ecc. (versione accettata dall'editore)
Dimensione
563.89 kB
Formato
Adobe PDF
|
563.89 kB | Adobe PDF | Visualizza/Apri |
Termine2021_Chapter_ModellingAccuracyAndTrustworth.pdf
accesso riservato
Tipologia:
Publisher's version/PDF
Dimensione
300.04 kB
Formato
Adobe PDF
|
300.04 kB | Adobe PDF | Visualizza/Apri Richiedi una copia |
Pubblicazioni consigliate
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.