Combination of confidence measures in isolated word recognition

Abstract

In the context of command-and-control applications, we exploit confidence measures in order to classify single-word utterances into two categories: utterances within the vocabulary which are recognized correctly, and other utterances, namely out-ofvocabulary (OOV) or misrecognized utterances. To this end, we investigate the classification error rate (CER) of several classes of confidence measures and transformations. In particular, we employed data-independent and data-dependent measures. The transformations we investigated include mapping to single confidence measures, LDA-transformed measures, and other linear combinations of these measures. These combinations are computed by means of neural networks trained with Bayesoptimal, and with Gardner-Derrida-optimal criteria. Compared to a recognition system without confidence measures, the selection of (various combinations of) confidence measures, the selection of suitable neural network architectures and training methods, continuously improves the CER. Additionally, we found that a linear perceptron generalizes better than a non-linear backpropagation network.