Robust Explanations Through Uncertainty Decomposition: A Path to Trustworthier AI

2025/07/17 by Chenrui Zhu, Zhu, Chenrui, Louenas Bounia +7
Computer Science · #Adversarial Robustness in Machine Learning #Explainable Artificial Intelligence (XAI) #FOS: Computer and information sciences #Machine Learning (cs.LG)

paper · pdf · doi:10.48550/arxiv.2507.12913

openalex publication_date 2025/07/17 · openalex created_date 2025/10/18 · openalex updated_date 2026/07/28

Abstract

Recent advancements in machine learning have emphasized the need for transparency in model predictions, particularly as interpretability diminishes when using increasingly complex architectures. In this paper, we propose leveraging prediction uncertainty as a complementary approach to classical explainability methods. Specifically, we distinguish between aleatoric (data-related) and epistemic (model-related) uncertainty to guide the selection of appropriate explanations. Epistemic uncertainty serves as a rejection criterion for unreliable explanations and, in itself, provides insight into insufficient training (a new form of explanation). Aleatoric uncertainty informs the choice between feature-importance explanations and counterfactual explanations. This leverages a framework of explainability methods driven by uncertainty quantification and disentanglement. Our experiments demonstrate the impact of this uncertainty-aware approach on the robustness and attainability of explanations in both traditional machine learning and deep learning scenarios.

Citations

• Reducing Aleatoric and Epistemic Uncertainty through Multi-modal Data Acquisition
• REPEAT: Improving Uncertainty Estimation in Representation Learning Explainability
• Robust Counterfactual Explanations in Machine Learning: A Survey
• A review of uncertainty quantification in medical image analysis: probabilistic and non-probabilistic methods
• Explainability through uncertainty: Trustworthy decision-making with neural networks
• Explainability is NOT a Game
• Logic-Based Explainability in Machine Learning
• SAFARI: Versatile and Efficient Evaluations for Robustness of Interpretability
• Out-of-Distribution Detection with Deep Nearest Neighbors
• Prior and Posterior Networks: A Survey on Evidential Deep Learning Methods For Uncertainty Estimation
• Trading Complexity for Sparsity in Random Forest Explanations
• How to choose an Explainability Method? Towards a Methodical Implementation of XAI in Practice
• Energy-based Out-of-distribution Detection
• Reliable Post hoc Explanations: Modeling Uncertainty in Explainability
• Opportunities and Challenges in Explainable Artificial Intelligence (XAI): A Survey
• Posterior Network: Uncertainty Estimation without OOD Samples via Density-Based Pseudo-Counts
• On The Reasons Behind Decisions
• Aleatoric and Epistemic Uncertainty with Random Forests
• The many Shapley values for model explanation
• DropConnect Is Effective in Modeling Uncertainty of Bayesian Deep\n Networks
• Reverse KL-Divergence Training of Prior Networks: Improved Uncertainty and Adversarial Robustness
• Explainable AI for Trees: From Local Explanations to Global Understanding
• Abduction-Based Explanations for Machine Learning Models
• On the Robustness of Interpretability Methods
• Evidential Deep Learning to Quantify Classification Uncertainty
• A Symbolic Approach to Explaining Bayesian Network Classifiers
• Predictive Uncertainty Estimation via Prior Networks
• Explanations based on the Missing: Towards Contrastive Explanations with Pertinent Negatives
• Explanation in Artificial Intelligence: Insights from the Social\n Sciences
• Explanation in artificial intelligence: Insights from the social sciences
• A Unified Approach to Interpreting Model Predictions
• What Uncertainties Do We Need in Bayesian Deep Learning for Computer Vision?
• Axiomatic Attribution for Deep Networks
• Simple and Scalable Predictive Uncertainty Estimation using Deep Ensembles
• "Why Should I Trust You?": Explaining the Predictions of Any Classifier
• Gradient-based learning applied to document recognition

Related

• Ensured: Explanations for Decreasing the Epistemic Uncertainty in Predictions
• Uncertainty Propagation in XAI: A Comparison of Analytical and Empirical Estimators
• Concerning Uncertainty -- A Systematic Survey of Uncertainty-Aware XAI
• Investigating the Impact of Model Instability on Explanations and Uncertainty
• Getting a CLUE: A Method for Explaining Uncertainty Estimates