New Bayesian Explanation Method Enhances Power Quality Disturbance Classifier Reliability

This paper introduces a novel post-hoc Bayesian explanation method for deep learning classifiers used in power quality disturbance recognition. The method employs a Laplace approximation to efficiently derive an approximate parameter posterior, enabling the generation of a distribution over disturbance-localization maps. This approach provides explanations with reliability measures, allowing for sharpened localization of specific events and indicating explanation dispersion under noise or domain shift. AI

IMPACT Enhances the reliability and interpretability of deployed AI models in critical infrastructure applications.

RANK_REASON The item is an academic paper detailing a new method for explaining machine learning models. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• Bayesian explanation
• Deep Ensembles
• Deep Learning Classifiers for Automated Detection of Gonioscopic Angle Closure Based on Anterior Segment OCT Images
• Laplace Approximation
• Monte Carlo Dropout
• POWER QUALITY DISTURBANCE CLASSIFICATION USING S-TRANSFORM AND RADIAL BASIS NETWORK
• Shap
• Yinsong Chen

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →