Confidence Estimation via Auxiliary Models
This work addresses the critical need for reliable confidence quantification in deep neural networks, particularly for safety-critical applications, by proposing a new, more effective confidence metric and a method to learn it.
This paper introduces True Class Probability (TCP) as a novel confidence criterion for deep neural classifiers, demonstrating its superiority over standard Maximum Class Probability (MCP). They propose an auxiliary model to learn TCP from data, which significantly outperforms strong baselines in failure prediction and self-training for domain adaptation across various datasets and architectures.
Reliably quantifying the confidence of deep neural classifiers is a challenging yet fundamental requirement for deploying such models in safety-critical applications. In this paper, we introduce a novel target criterion for model confidence, namely the true class probability (TCP). We show that TCP offers better properties for confidence estimation than standard maximum class probability (MCP). Since the true class is by essence unknown at test time, we propose to learn TCP criterion from data with an auxiliary model, introducing a specific learning scheme adapted to this context. We evaluate our approach on the task of failure prediction and of self-training with pseudo-labels for domain adaptation, which both necessitate effective confidence estimates. Extensive experiments are conducted for validating the relevance of the proposed approach in each task. We study various network architectures and experiment with small and large datasets for image classification and semantic segmentation. In every tested benchmark, our approach outperforms strong baselines.