Error-correction detection and response generation in a spoken dialogue system

Bulyko, Ivan; Kirchhoff, Katrin; Ostendorf, Mari; Goldberg, Jack

doi:10.1016/j.specom.2004.09.009

Cited by 28 publications

(24 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In contrast, neutral turns in the FAH dimension are rejected less than expected. Surprisingly, FrAng does not interact with AsrMis as observed in (Bulyko et al, 2005) but they use the full word error rate measure instead of the binary version used in this paper. Table 3: FAH -REJ interaction.…”

Section: Within Turn Interactionsmentioning

confidence: 99%

“…We perform our analysis on three high level dialogue factors: frustration/anger, certainty and correctness. Frustration and anger have been observed as the most frequent emotional class in many dialogue systems (Ang et al, 2002) and are associated with a higher word error rate (Bulyko et al, 2005). For this reason, we use the presence of emotions like frustration and anger as our first dialogue factor.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Dependencies between student state and speech recognition problems in spoken tutoring dialogues

Rotaru

Litman

2006

Proceedings of the 21st International Conference on Computational Linguistics and the 44th Annual Meeting of the ACL - ACL '06

View full text Add to dashboard Cite

Speech recognition problems are a reality in current spoken dialogue systems. In order to better understand these phenomena, we study dependencies between speech recognition problems and several higher level dialogue factors that define our notion of student state: frustration/anger, certainty and correctness. We apply Chi Square (χ2) analysis to a corpus of speech-based computer tutoring dialogues to discover these dependencies both within and across turns. Significant dependencies are combined to produce interesting insights regarding speech recognition problems and to propose new strategies for handling these problems. We also find that tutoring, as a new domain for speech applications, exhibits interesting tradeoffs and new factors to consider for spoken dialogue design.

show abstract

Section: Within Turn Interactionsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Dependencies between student state and speech recognition problems in spoken tutoring dialogues

Rotaru

Litman

2006

Proceedings of the 21st International Conference on Computational Linguistics and the 44th Annual Meeting of the ACL - ACL '06

View full text Add to dashboard Cite

show abstract

“…It is important to point out that we have not allowed barge-in (the speech recognizer is inactive while the system is speaking). This makes for a less flexible dialogue than may be desirable, but in certain situations such as recognition error spirals [12] it may be advisable not to allow the user to interrupt while the system is trying to reach a stable, mutually understood dialogue state, especially if the user's perception of reliability in identity authentication rests partly on how much under control the dialogue is seen to be.…”

Section: Turn Managementmentioning

confidence: 99%

Embodied conversational agents for voice-biometric interfaces

Hernández-Trapote

López-Mencía

Díaz

et al. 2008

Proceedings of the 10th International Conference on Multimodal Interfaces

View full text Add to dashboard Cite

In this article we present a research scheme which aims to analyze the use of Embodied Conversational Agent (ECA) technology to improve the robustness and acceptability of speaker enrolment and verification dialogues designed to provide secure access through natural and intuitive speaker recognition. In order to find out the possible effects of the visual information channel provided by the ECA, tests were carried out in which users were divided into two groups, each interacting with a different interface (metaphor): an ECA Metaphor group -with an ECA-, and a VOICE Metaphor group -without an ECA-. Our evaluation methodology is based on the ITU-T P.851 recommendation for spoken dialogue system evaluation, which we have complemented to cover particular aspects with regard to the two major extra elements we have incorporated: secure access and an ECA. Our results suggest that likeability-type factors and system capabilities are perceived more positively by the ECA metaphor users than by the VOICE metaphor users. However, the ECA's presence seems to intensify users' privacy concerns.

show abstract

“…The authors in [8] proposed a system which integrates an error correction detection module with a modified dialogue strategy. In the study [9], a machine-learning approach employed automatically derived prosodic features, the speech recognition process, experimental conditions and the dialogue history to identify user corrections of speech recognition errors.…”

Section: Introductionmentioning

confidence: 99%

Classification of Correction Turns in Multilingual Dialogue Corpus

Kraljevski

Hirschfeld²

2018

Interspeech 2018

View full text Add to dashboard Cite

This paper presents a multiclass classification of correction dialog turns using machine learning. The classes are determined by the type of the introduced recognition errors while performing WOz trials and creating the multilingual corpus. Three datasets were obtained using different sets of acoustic-prosodic features on the multilingual dialogue corpus.The classification experiments were done using different machine learning paradigms: Decision Trees, Support Vector Machines and Deep Learning. After careful experiments setup and optimization on the hyper-parameter space, the obtained classification results were analyzed and compared in the terms of accuracy, precision, recall and F1 score. The achieved results are comparable with those obtained in similar experiments on different tasks and speech databases.

show abstract

Error-correction detection and response generation in a spoken dialogue system

Cited by 28 publications

References 21 publications

Dependencies between student state and speech recognition problems in spoken tutoring dialogues

Dependencies between student state and speech recognition problems in spoken tutoring dialogues

Embodied conversational agents for voice-biometric interfaces

Classification of Correction Turns in Multilingual Dialogue Corpus

Contact Info

Product

Resources

About