This work shows how a dialogue model can be represented as a Partially Observable Markov Decision Process (POMDP) with observations composed of a discrete and continuous component. The continuous component enables the model to directly incorporate a confidence score for automated planning. Using a testbed simulated dialogue management problem, we show how recent optimization techniques are able to find a policy for this continuous POMDP which outperforms a traditional MDP approach. Further, we present a method for automatically improving handcrafted dialogue managers by incorporating POMDP belief state monitoring, including confidence score information. Experiments on the testbed system show significant improvements for several example handcrafted dialogue managers across a range of operating conditions. 1

Abstract not found

Abstract not found

In this paper two studies are presented in which the detection of speech recognition errors on the word level was examined.

This article focuses on the analysis and prediction of corrections, defined as turns where a user tries to correct a prior error made by a spoken dialogue system. We describe our labeling procedure of various corrections types and statistical analyses of their features in a corpus collected from a train information spoken dialogue system. We then present results of machinelearning experiments designed to identify user corrections of speech recognition errors. We investigate the predictive power of features automatically computable from the prosody of the turn, the speech recognition process, experimental conditions, and the dialogue history. Our best-performing features reduce classification error from baselines of 25.70–28.99 % to 15.72%. 1.

The problems caused by imperfect speech recognition in spoken dialogue systems are well known: they confound the ability of the system to manage the dialogue, and can lead to both user frustration and task failure. Speech recognition errors are likely to persist for the foreseeable future, and so the development and adoption of a well-founded approach to the handling of error situations may be an important component in achieving general public acceptability for systems of this kind. In this paper, we compare two studies of user behaviour in response to speech recognition errors in quite different dialog applications; the analysis supports the view that user behaviour during error conditions contains a large component that is independent of the domain of the dialogue. The prospect of a consistent response to errors across a wide range of domains enhances the prospects for a general theory of error recognition and repair. 1

Most dialog systems explicitly confirm user-provided task-relevant concepts. User responses to these system confirmations (e.g. corrections, topic changes) may be misrecognized because they contain unrequested task-related concepts. In this paper, we propose a concept-specific language model adaptation strategy where the language model (LM) is adapted to the concept type(s) actually present in the user’s post-confirmation utterance. We evaluate concept type classification and LM adaptation for post-confirmation utterances in the Let’s Go! dialog system. We achieve 93 % accuracy on concept type classification using acoustic, lexical and dialog history features. We also show that the use of concept type classification for LM adaptation can lead to improvements in speech recognition performance. 1

are those of the author and should not be interpreted as representing the official policies, either express or implied, of any sponsoring institution, the U.S. government, or any other entity. Keywords: spoken dialog systems, conversational spoken language interfaces, error detection, error recovery strategies, error recovery policies, dialog management, RavenClaw, implicitly-supervised One of the most important and persistent problems in the development of conversational spoken language interfaces is their lack of robustness when confronted with understanding-errors. Most of these errors stem from limitations in current speech recognition technology, and, as a result, appear across all domains and interaction types. There are two approaches towards increased robustness: prevent the errors from happening, or recover from them through conversation, by interacting with the users. In this dissertation we have engaged in a research program centered on the second approach. We argue that three capabilities are needed in order to seamlessly and efficiently recover from errors: (1) systems must be able to detect the errors, preferably as soon as they happen, (2) systems must be equipped with a rich repertoire of error recovery strategies that can be used to set the conversation back on track, and (3) systems must know how to choose optimally between different recovery

This document is available from http://ilk.kub.nl/downloads/pub/papers/ilk0210.ps.gz. All rights reserved Induction of Linguistic Knowledge, Tilburg University and CNTS Research Group, University of Antwerp

For spoken dialogue systems, errors can occur on different levels of the system's architecture. One of the principal causes for errors during a dialogue session are erroneous recognition results which often lead to incorrect semantic interpretations. Even if the speech input signal has been correctly recognized, a natural language understanding component can produce errorprone sentence meanings due to the limitations of its underlying model. To cope with this problem, we introduce a multi-level error-detection mechanism based on several features in order to find erroneous recognitions, error-prone semantic interpretations as well as ambiguities, and contradictions. Here, the confidence output of one level directly serves as an additional input for the subsequent level. The proposed features and scoring criteria are passed to the dialogue manager which then determines the subsequent dialogue action.