Learning hidden unit contributions for unsupervised speaker adaptation of neural network acoustic models

Abstract

This paper proposes a simple yet effective model-based neural network speaker adaptation technique that learns speakerspecific hidden unit contributions given adaptation data, without requiring any form of speaker-adaptive training, or labelled adaptation data. An additional amplitude parameter is defined for each hidden unit; the amplitude parameters are tied for each speaker, and are learned using unsupervised adaptation. We conducted experiments on the TED talks data, as used in the International Workshop on Spoken Language Translation (IWSLT) evaluations. Our results indicate that the approach can reduce word error rates on standard IWSLT test sets by about 8-15% relative compared to unadapted systems, with a further reduction of 4-6% relative when combined with feature-space maximum likelihood linear regression (fMLLR). The approach can be employed in most existing feed-forward neural network architectures, and we report results using various hidden unit activation functions: sigmoid, maxout, and rectifying linear units (ReLU).

An analysis of environment, microphone and data simulation mismatches in robust speech recognition

“…In most corpora, the training speakers differ from the test speakers. This is widely recognized as good practice and many solutions are available to improve robustness to this mismatch (Gales, 1998;Shinoda, 2011;Karafiát et al, 2011;Swietojanski and Renals, 2014). By contrast, the acoustic conditions of the training data often match (or cover) those of the test data.…”

An Investigation of Deep Neural Networks for Multilingual Speech Recognition Training and Adaptation

“…Learning hidden unit contribution (LHUC) is a method that linearly re-combines hidden units in a speaker-or environmentdependent manner [14,25]. Given adaptation data, LHUC rescales the contributions (amplitudes) of the hidden units in the model without actually modifying their feature receptors.…”

“…Similar approaches have been proposed independently in [12] and [13]. Researchers have also introduced learning hidden unit contribution (LHUC) to weight hidden unit activations in a speaker-or environment-dependent manner [14]. It was shown that LHUC results in consistent WER reductions for speaker and environment adaptation [15].…”

An assembly system based on industrial robot with binocular stereo vision

“…In this paper, a deep multilayer neural network is used for the object recognition [12][13][14]. Its learning rule is to apply the steepest descend method to adjust the weights and thresholds of the neural network according to the minimum sum of the square error.…”