Uncertainty Gating for Cost-Aware Explainable Artificial Intelligence
This work addresses the cost and reliability issues in explainable AI for users needing interpretable predictions, but it is incremental as it builds on existing uncertainty and explanation methods.
The paper tackled the problem of computationally expensive and unreliable post-hoc explanation methods in AI by proposing epistemic uncertainty as a low-cost proxy for explanation reliability, showing a strong negative correlation between epistemic uncertainty and explanation stability across multiple datasets and architectures.
Post-hoc explanation methods are widely used to interpret black-box predictions, but their generation is often computationally expensive and their reliability is not guaranteed. We propose epistemic uncertainty as a low-cost proxy for explanation reliability: high epistemic uncertainty identifies regions where the decision boundary is poorly defined and where explanations become unstable and unfaithful. This insight enables two complementary use cases: `improving worst-case explanations' (routing samples to cheap or expensive XAI methods based on expected explanation reliability), and `recalling high-quality explanations' (deferring explanation generation for uncertain samples under constrained budget). Across four tabular datasets, five diverse architectures, and four XAI methods, we observe a strong negative correlation between epistemic uncertainty and explanation stability. Further analysis shows that epistemic uncertainty distinguishes not only stable from unstable explanations, but also faithful from unfaithful ones. Experiments on image classification confirm that our findings generalize beyond tabular data.