Coalitional Bayesian autoencoders: Towards explainable unsupervised deep learning with applications to condition monitoring under covariate shift
Accepted version
Peer-reviewed
Repository URI
Repository DOI
Change log
Authors
Brintrup, A
Abstract
This paper aims to improve the explainability of autoencoder (AE) pre- dictions by proposing two novel explanation methods based on the mean and epistemic uncertainty of log-likelihood estimates, which naturally arise from the probabilistic formulation of the AE, the Bayesian autoencoder (BAE). These formulations contrast the conventional post-hoc explanation methods for AEs, which incur additional modelling effort and implementations. We further extend the methods for sensor-based explanations, aggregating the explanations at the sensor level instead of the lower feature level.
Description
Keywords
Explainable deep learning, Bayesian autoencoders, Condition monitoring, Industry 4, 0, Internet of Things
Journal Title
Applied Soft Computing
Conference Name
Journal ISSN
1568-4946
1872-9681
1872-9681
Volume Title
Publisher
Elsevier BV