Coalitional Bayesian Autoencoders -- Towards explainable unsupervised deep learning
arXiv:2110.10038
Abstract
This paper aims to improve the explainability of Autoencoder's (AE) predictions by proposing two explanation methods based on the mean and epistemic uncertainty of log-likelihood estimate, which naturally arise from the probabilistic formulation of the AE called Bayesian Autoencoders (BAE). To quantitatively evaluate the performance of explanation methods, we test them in sensor network applications, and propose three metrics based on covariate shift of sensors : (1) G-mean of Spearman drift coefficients, (2) G-mean of sensitivity-specificity of explanation ranking and (3) sensor explanation quality index (SEQI) which combines the two aforementioned metrics. Surprisingly, we find that explanations of BAE's predictions suffer from high correlation resulting in misleading explanations. To alleviate this, a "Coalitional BAE" is proposed, which is inspired by agent-based system theory. Our comprehensive experiments on publicly available condition monitoring datasets demonstrate the improved quality of explanations using the Coalitional BAE.
arXiv commentPreprint submitted to Journal of Applied Soft Computing
References in corpus (10)
- Scikit-learn: Machine Learning in Python
- Deep learning for time series classification: a review
- Axiomatic Attribution for Deep Networks
- Learning Important Features Through Propagating Activation Differences
- Machine Learning and Deep Learning Algorithms for Bearing Fault Diagnostics -- A Comprehensive Review
- A Benchmark for Interpretability Methods in Deep Neural Networks
- Explaining Anomalies Detected by Autoencoders Using SHAP
- Quality of Uncertainty Quantification for Bayesian Neural Network Inference
- Bayesian Autoencoders for Drift Detection in Industrial Environments
- Bayesian Autoencoders: Analysing and Fixing the Bernoulli likelihood for Out-of-Distribution Detection