We describe the design and evaluation of two different dynamic student uncertainty adaptations in wizarded versions of a spoken dialogue tutoring system. The two adaptive systems adapt to each student turn based on its uncertainty, after an unseen human “wizard” performs speech recognition and natural language understanding and annotates the turn for uncertainty. The design of our two uncertain...

Background: Speech perception ability depends on auditory and extra-auditory elements. The signal-to-noise ratio (SNR) is an extra-auditory element that has an effect on the ability to normally follow speech and maintain a conversation. Speech in noise perception difficulty is a common complaint of the elderly. In this study, the importance of SNR magnitude as an extra-auditory effect on speech...

While it is often fairly straightforward to estimate the reliability of speech features in the time-frequency domain, this may not be true in other domains more amenable to speech recognition, such as e.g. for RASTA-PLP features or those obtained with the ETSI advanced front end. In such cases, one useful approach is to estimate the uncertainties in the domain where noise-reduction preprocessin...

This study deals with uncertainty and attention in audiovisual speech perception. Subjects were exposed to audiovisual stimuli where independent variable was bivalent: either blocked to one ear or uncertainty about in which ear the next stimulus would appear. The hypothesis, that uncertainty would inhibit audiovisual integration was motivated by an earlier study (Öhrström et al, 2011) but these...

Most conventional techniques for noise adaptation assume a clean initial speech model which is adapted to a specific noise condition using adaptation data accumulated from the condition. In this paper, a different problem is considered, i.e. adapting a noisy speech model to a specific noise condition. For example, the initial noisy model may be a multi-condition model which is used to provide m...

In this paper, the uncertainty is defined as the mean square error between a given enhanced noisy observation vector and the corresponding clean one. Then, a DNN is trained by using enhanced noisy observation vectors as input and the uncertainty as output with a training database. In testing, the DNN receives an enhanced noisy observation vector and delivers the estimated uncertainty. This unce...

Many techniques in speech processing require inference based on observations that are often noisy, incomplete or scarce. In such situations, it is necessary to draw on statistical techniques that themselves must be robust to the nature of the observations. The Bayesian method is a school of thought within statistics that provides such a robust framework for handling “difficult” data. In particu...