Augmentation by Counterfactual Explanation -- Fixing an Overconfident Classifier
This addresses the issue of overconfidence in AI models for high-stakes domains like healthcare and autonomous driving, though it is incremental as it builds on existing counterfactual explanation techniques.
The paper tackles the problem of overconfident classifiers in critical applications by fine-tuning a pre-trained model using augmentations from a counterfactual explainer (ACE) to improve uncertainty measures while maintaining predictive performance. The results show improved uncertainty detection for far-OOD, near-OOD, and ambiguous samples, with performance competitive to state-of-the-art methods.
A highly accurate but overconfident model is ill-suited for deployment in critical applications such as healthcare and autonomous driving. The classification outcome should reflect a high uncertainty on ambiguous in-distribution samples that lie close to the decision boundary. The model should also refrain from making overconfident decisions on samples that lie far outside its training distribution, far-out-of-distribution (far-OOD), or on unseen samples from novel classes that lie near its training distribution (near-OOD). This paper proposes an application of counterfactual explanations in fixing an over-confident classifier. Specifically, we propose to fine-tune a given pre-trained classifier using augmentations from a counterfactual explainer (ACE) to fix its uncertainty characteristics while retaining its predictive performance. We perform extensive experiments with detecting far-OOD, near-OOD, and ambiguous samples. Our empirical results show that the revised model have improved uncertainty measures, and its performance is competitive to the state-of-the-art methods.