Distance-based Confidence Score for Neural Network Classifiers
This addresses the need for reliable confidence measurement in classifiers, which is crucial for many applications, though it is an incremental improvement over existing methods.
The paper tackles the problem of quantifying prediction confidence in neural network classifiers by proposing a simple, scalable method based on data embeddings from the penultimate layer, showing significant improvement over traditional confidence scores in tasks like classification error prediction, ensemble weighting, and novelty detection.
The reliable measurement of confidence in classifiers' predictions is very important for many applications and is, therefore, an important part of classifier design. Yet, although deep learning has received tremendous attention in recent years, not much progress has been made in quantifying the prediction confidence of neural network classifiers. Bayesian models offer a mathematically grounded framework to reason about model uncertainty, but usually come with prohibitive computational costs. In this paper we propose a simple, scalable method to achieve a reliable confidence score, based on the data embedding derived from the penultimate layer of the network. We investigate two ways to achieve desirable embeddings, by using either a distance-based loss or Adversarial Training. We then test the benefits of our method when used for classification error prediction, weighting an ensemble of classifiers, and novelty detection. In all tasks we show significant improvement over traditional, commonly used confidence scores.