Global Explanations with Decision Rules: a Co-learning Approach - Archive ouverte HAL Accéder directement au contenu
Article Dans Une Revue Proceedings of Machine Learning Research Année : 2021

Global Explanations with Decision Rules: a Co-learning Approach

Géraldin Nanfack
  • Fonction : Auteur
  • PersonId : 1182288
Paul Temple
Benoît Frenay
  • Fonction : Auteur
  • PersonId : 1033494

Résumé

Black-box machine learning models can be extremely accurate. Yet, in critical applications such as in healthcare or justice, if models cannot be explained, domain experts will be reluctant to use them. A common way to explain a black-box model is to approximate it by a simpler model such as a decision tree. In this paper, we propose a co-learning framework to learn decision rules as explanations of black-box models through knowledge distillation and simultaneously constrain the blackbox model by these explanations; all of this in a differentiable manner. To do so, we introduce the soft truncated Gaussian mixture analysis (STruGMA), a probabilistic model which encapsulates hyper-rectangle decision rules. With STruGMA, global explanations can be provided by any rule learner such as decision lists, sets or trees. We provide evidences through experiments that our framework can globally explain black-box models such as neural networks. In particular, the explanation fidelity is increased, while the accuracy of the models is marginally impacted.
Fichier principal
Vignette du fichier
nanfack21a.pdf (1.44 Mo) Télécharger le fichier
Origine : Fichiers éditeurs autorisés sur une archive ouverte

Dates et versions

hal-03840082 , version 1 (04-11-2022)

Identifiants

  • HAL Id : hal-03840082 , version 1

Citer

Géraldin Nanfack, Paul Temple, Benoît Frenay. Global Explanations with Decision Rules: a Co-learning Approach. Proceedings of Machine Learning Research, 2021, Uncertainty in Artificial Intelligence, 27-30 July 2021, Online, 161, pp.589-599. ⟨hal-03840082⟩
6 Consultations
8 Téléchargements

Partager

Gmail Facebook X LinkedIn More