McNair, A. and Waibel, A. (1994). Improving recognizer acceptance through robust, natural speech repair. In Proc. ICSLP'94, pages 1299--1302, Yokohama, Japan. Home/Search Document Not in Database Summary Related Articles Check |
This paper is cited in the following contexts:
Detection and Transcription of OOV Words - Fetter (1998) (3 citations) (Correct)
....the speech recognizer is a transparent module of a larger application, for example the Verbmobil speech to speech translation system, the situation is more difficult. Whereas in vocabulary recognition errors can often be repaired (by repeating the utterance and speaking more clearly, for instance) [McNair and Waibel, 1994], out of vocabulary errors cannot. The problem is that the user cannot differentiate between in vocabulary and out of vocabulary errors because they do not have access to the recognizer hypotheses. Moreover, since most naive users are likely to ignore the vocabulary limitations of the system, they ....
McNair, A. and Waibel, A. (1994). Improving recognizer acceptance through robust, natural speech repair. In Proc. ICSLP'94, pages 1299--1302, Yokohama, Japan.
Interactive Translation of Conversational Speech - Waibel (1996) (30 citations) Self-citation (Waibel) (Correct)
....as the user most likely does not know if an output translation in an unknown language is correct or not. Semantic representations in natural language processing have, of course, been studied extensively over the years, leading to a number of Interlingua based text translation systems (see [11][12] 13] for review) We find the use of an interlingua based approach particularly advantageous for the translation of spontaneous speech, as spoken language is syntactically more ill formed and less reliable, but the semantics typically more contained. For each recognition hypothesis emerging ....
....recovery mechanisms, the translation station also elicits somewhat more benign user speaking style than human to human conversational speech. Work is in progress that exploits this opportunity for error correction. To recover from human and machine error, a number of strategies have been explored [11], including repair by respeaking, spelling, and handwriting as alternative redundant modes of human computer interaction. Recovery can typically be achieved within one or two tries. The JANUS Interactive Translation of Conversational Speech 14 II system also offers other simple forms of ....
A.E.McNair and A.Waibel, "Improving Recognizer Acceptance through Robust, Natural Speech Repair", ICSLP 1994, Vol. 3, pp. 1299.
Integrating Spelling Into Spoken Dialogue Recognition - Hild, Waibel (1995) Self-citation (Waibel) (Correct)
.... 21 September 1995 1. INTRODUCTION Applications of spelling in spontaneous speech include the recognition of spelled names or addresses, as well as repair dialogues, where spelling can be used to disambiguate the inevitable recognition errors or to introduce out of vocabulary words to the system [4]. All these applications have in common that in natural speech, sentences carrying spellings will rarely consist of plain, clean letter sequences. Rather, the spelling is embedded in spontaneous speech, which introduces, corrects, comments on, paraphrases or further specifies the individual ....
Arthur E. McNair and Alex Waibel. Improving Recognizer Acceptance Through Robust, Natural Speech Repair. In International Conference on Speech and Language Processing, pages S22--15.1, 1994.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC