In eXplainable Artificial Intelligence (XAI), several counterfactual explainers have been proposed, each focusing on some desirable properties of counterfactual instances: minimality, actionability, stability, diversity, plausibility, discriminative power. We propose an ensemble of counterfactual explainers that boosts weak explainers, which provide only a subset of such properties, to a powerful method covering all of them. The ensemble runs weak explainers on a sample of instances and of features, and it combines their results by exploiting a diversity-driven selection function. The method is model-agnostic and, through a wrapping approach based on autoencoders, it is also data-agnostic.
Ensemble of counterfactual explainers
Guidotti R.;Ruggieri S.
2021
Abstract
In eXplainable Artificial Intelligence (XAI), several counterfactual explainers have been proposed, each focusing on some desirable properties of counterfactual instances: minimality, actionability, stability, diversity, plausibility, discriminative power. We propose an ensemble of counterfactual explainers that boosts weak explainers, which provide only a subset of such properties, to a powerful method covering all of them. The ensemble runs weak explainers on a sample of instances and of features, and it combines their results by exploiting a diversity-driven selection function. The method is model-agnostic and, through a wrapping approach based on autoencoders, it is also data-agnostic.| File | Dimensione | Formato | |
|---|---|---|---|
|
prod_460372-doc_179466.pdf
solo utenti autorizzati
Descrizione: Ensemble of Counterfactual Explainers
Tipologia:
Versione Editoriale (PDF)
Licenza:
NON PUBBLICO - Accesso privato/ristretto
Dimensione
1.2 MB
Formato
Adobe PDF
|
1.2 MB | Adobe PDF | Visualizza/Apri Richiedi una copia |
|
2308.15194v1.pdf
accesso aperto
Descrizione: This is the Submitted version (preprint) of the following paper: Guidotti R., Ruggieri S. “Ensemble of Counterfactual Explainers”, submitted to “Discovery Science. 24th International Conference, DS 2021”, Halifax, NS, Canada, October 11–13, 2021. The final published version is available on the publisher’s website https://link.springer.com/book/10.1007/978-3-030-88942-5.
Tipologia:
Documento in Pre-print
Licenza:
Creative commons
Dimensione
2.32 MB
Formato
Adobe PDF
|
2.32 MB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
