Getting a CLUE: A Method for Explaining Uncertainty Estimates

By Javier Antorán, Umang Bhatt, Tameem Adel, Adrian Weller, José Miguel Hernández-Lobato. 2020

135 citations recorded by the source corpus.

Both uncertainty estimation and interpretability are important factors for trustworthy machine learning systems. However, there is little work at the intersection of these two areas. We address this gap by proposing a novel method for interpreting uncertainty estimates from differentiable probabilistic models, like Bayesian Neural Networks (BNNs). Our method, Counterfactual Latent Uncertainty Explanations (CLUE), indicates how to change an input, while keeping it on the data manifold, such that a BNN becomes more confident about the input's prediction. We validate CLUE through 1) a novel framework for evaluating counterfactual explanations of uncertainty, 2) a series of ablation experiments, and 3) a user study. Our experiments show that CLUE outperforms baselines and enables practitioners to better understand which input patterns are responsible for predictive uncertainty.

Subfields: Artificial Intelligence, Safety, Robustness & Interpretability

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Safety, Robustness & Interpretability

Citations in the corpus

Interpretable Machine Learning: Fundamental Principles and 10 Grand Challenges (2021, background): “Tomsett et al. (2020) discusses the importance of uncertainty quantification and interpretability, and Antoràn et al. (2020) discusses interpretable uncertainty estimation.”
Counterfactual Explanations and Algorithmic Recourses for Machine Learning: A Review (2020, unknown): “Antoran et al. [15] use CFEs to find changes in a datapoint that would help a classifier have a higher confidence in its prediction.”
Explaining in Style: Training a GAN to explain a classifier in StyleSpace (2021, background): “In [31, 7, 32, 1] generative counterfactual explanations are produced, yet their visualization changes all relevant attributes at once, as shown in Fig.”
A survey of uncertainty in deep neural networks (2021, background): “[342] claimed to have published the first work on explainable uncertainty estimation.”
Interpretable deep learning: interpretation, interpretability, trustworthiness, and beyond (2021, methodology): “An approach to evaluate the algorithm of counterfactual examples [11] was proposed, where a user-study experiment was used to validate their approaches.”
Deep Neural Networks and Tabular Data: A Survey (2021, background): “To alleviate this problem and to suggest realistic looking counterfactuals, researchers have suggested building recourse suggestions on generative models [179]–[184].”
Sample-Efficient Optimization in the Latent Space of Deep Generative Models via Weighted Retraining (2020, background): “automatic machine learning [41, 42, 76], conditional image generation [48, 47], and model explainability [2].”
Uncertainty as a Form of Transparency: Measuring, Communicating, and Using Uncertainty (2020, methodology): “[2] introduced CLUE, a method to help identify which input features are responsible for models’ uncertainty.”