Zue, V., J. Glass, D. Goodine, H. Leung, M. Phillips and S. Seneft, "The VOY- AGER Speech Understanding system: Preliminary Development and Evaluation," in Proc. ICASSP-90, pp. 73-76, Albuquerque, NM, April 1990.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....systems. A lot of online dialogue systems have been described mainly for dealing with times or train schedules , for example [3, 18, 14, 10, 7] This is of little interest for the Ypa. However, dialogue systems have also been built for retrieving addresses from Yellow Pages, for example Voyager [24, 22, 6] and Galaxy, but the implementations are restricted to sample domains or small scale address databases (e.g. 150 objects in Voyager [6] or 2400 tourist related listings in Galaxy [23] This is of course not unrelated to the fact that these are spoken language dialogue systems. However, the task ....

Zue, V., Glass, J., Goodine, D., Leung, H., Phillips, M., Polifroni, J., and Seneff, S. The VOYAGER Speech Understanding System: Preliminary Development and Evaluation. In Proceedings of IEEE International Conference on Acoustics, Speech and Signal Processing (1990).

....concerned with either front end or back end, or less concerned with a particular application. Most natural language dialogue systems are concerned with the problem of schedules or time [1 5] something that does not apply to the YPA. More closely related to the YPA is the Voyager dialogue system [6 8] which deals with addresses from NYNEX Yellow Pages. However, as noted above, this project uses source data that is more than a pure address database. Hence it seems there are IR systems that are even more closely related. This paper uses concepts related to those in Strzalkowski [9] and ....

Zue V, Glass J, Goodine D, Leung H, Phillips M, Polifroni J and Seneff S: `The VOYAGER speech understanding system: preliminary development and evaluation', Proceedings of IEEE International Conference on Acoustics, Speech and Signal Processing (1990).

....with the problem of schedules or time (among others: Aust et al. 1995) Wahlster, 1993) Sikorski and Allen, 1996) McGlashan et al. 1992; Heisterkamp, McGlashan, and Youd, 1992) something that does not apply to the Ypa. More closely related to the Ypa is the Voyager dialogue system (Zue et al. 1990; Zue, 1994; Glass et al. 1995) which also deals with addresses from Yellow Pages (NYNEX) However, as noted above, this project uses source data that is more than a pure address database. Hence it seems there are IR systems that are even more closely related. This paper uses concepts related to ....

Zue, V., J. Glass, D. Goodine, H. Leung, M. Phillips, J. Polifroni, and S. Seneff. 1990. The VOYAGER Speech Understanding System: Preliminary Development and Evaluation. In Proceedings of IEEE International Conference on Acoustics, Speech and Signal Processing.

....The Berkeley Restaurant Project (BeRP) is a medium sized vocabulary, speaker independent speech understanding system whose domain is knowledge about restaurants in the city of Berkeley. BeRP is similar to other spontaneous speech understanding systems that have been developed recently (Price 1990; Zue et al. 1990). Its primary purpose is to serve as a testbed for many ideas relating to speech recognition and understanding, including robust acoustic processing, connectionist modeling, foreign accent detection and modeling, automatic induction of multiple pronunciation lexicons, and the tight coupling of ....

....rate greater than 100.0 . While these error rates are much higher than for tasks involving read speech (as opposed to spontaneous speech) they are comparable to the initial results obtained at the Massachusetts Institute of Technology for their spontaneous speech understanding system Voyager (Zue et al. 1990). These results are not as good as current state of the art speech understanding systems for a couple of reasons. First, we are using an unsmoothed bigram grammar which is fairly constraining and second, we are using monophone sub word models. Both of these could be improved, but this is outside ....

ZUE, V., J. GLASS, D. GOODINE, H. LEUNG, M. PHILLIPS, & S. SENEFF. 1990. The VOYAGER speech understanding system: Preliminary development and evaluation. In Proceedings Int'l Conference on Acoustics Speech and Signal Processing, 73--76, Albuquerque, New Mexico.

....were limited to the recognition of isolated words or discrete speech, to the recognition of small vocabularies, or to speaker dependent, carefully articulated and constrained inputs. Much of this research was also divorced from natural language understanding. However, with systems like VOYAGER (Zue et al. 1990), HEARSAY II (Erman et al. 1980) SPHINX (Lee et al. 1990) and MINDS (Young et al. 1989) significant advances in automatic speech recognition and natural language understanding are being made. The work of these researchers shows that integrating higher level knowledge sources, such as ....

V. Zue, J. Glass, D. Goodine, H. Leung, M. Phillips, J. Polifroni and S. Seneff (1990) "The VOYAGER Speech Understanding System: Preliminary Development and Evaluation," Proceedings of ICASSP 90, 73-76.

....Restaurant Project (BeRP) 4] BeRP is a medium sized vocabulary, speaker independent speech understanding system whose domain is knowledge about restaurants in the city of Berkeley. The BeRP system is similar to other spontaneous speech understanding systems that have been developed recently [5, 6]. One of the distinguishing characteristics of BeRP is that it uses a speech recognizer that combines neural networks and Hidden Markov Models (HMMs) The neural network is a Multi layer Perceptron (MLP) which is used to estimate the acoustic likelihoods for the HMMs [7] In the next sections we ....

V. Zue, J. Glass, D. Goodine, H. Leung, M. Phillips, and S. Seneff. The VOYAGER speech understanding system: Preliminary development and evaluation. In Proceedings Int'l Conference on Acoustics Speech and Signal Processing, pages 73--76, Albuquerque, New Mexico, 1990.

....[3, 4, 8, 9, 14, 24, 26, 27, 28, 33] For example, the waxholm system [4] gives information on boat traffic in the Stockholm archipelago. There are other systems for accessing time schedules for trains [8, 9, 26] the ATIS (Air Travel Information System) domain [27, 33] urban navigation [3, 35]; automobile classifieds [24] etc. 1.1.2 Multilingual Translation Projects The goal for multilingual translation systems is to enable humans to communicate with one another in their native tongues. To achieve this goal, speech recognition, machine translation, and speech synthesis technologies ....

....because the system only needs to deal with the diversity of the conversation from the human side, and the topic of the conversation is usually focused in the application domain. We have not found other similar systems reported in the literature so far, aside from galaxy s predecessor voyager [34, 35]. Voyager only has a city guide domain which is much simpler than the current city guide subdomain of galaxy, and it was not initially designed to easily support multiple languages. Through a trialand error process that involved several steps of redesign, it eventually could support three ....

V. Zue, J. Glass, D. Goodine, H. Leung, M. Phillips, J. Polifroni, and S. Seneff. The VOYAGER speech understanding system: Preliminary development and evaluation. In Proc. ICASSP, pages 73--76, 1990.

....rule based aspect of our hybrid approach. The procedure involves labelling a training corpus, writing a set of context free rules, and boot strapping with the nat ural language parser TINA [73] TINA has previously been used with the SUMMIT recognizer [97] to parse sentences in the VOYAGEI domain [98] for navigation, and the ATIS domain [66] for retrieving air travel information. The formalism of TINA derives a network from a context free grammar, and the connecting arcs in the network are associated with probabilities. When a sentence is parsed, a set of parse nodes are created and linked ....

Zue, V., J. Glass, D. Goodine, H. Leung, M. Phillips and S. Seneft, "The VOY- AGER Speech Understanding system: Preliminary Development and Evaluation," in Proc. ICASSP-90, pp. 73-76, Albuquerque, NM, April 1990.

....system. As a testbed for multi lingual research, a multi lingual information retrieval system is currently under development in the Spoken Language Systems group at MIT. This system, known as the multi lingual voyager system, is designed to provide travel information for the city of Cambridge [38, 39, 42]. Voyager currently has the capability to understand queries in either English or Japanese [7] and is being ported to French, Italian and German. Within the multi lingual voyager domain, ALI can be performed as a two step process. The first step is to perform a fast match to provide an ordered ....

Victor Zue, James Glass, David Goodine, Hong Leung, Michael Phillips, Joseph Polifroni, and Stephanie Seneff. The voyager speech understanding system: Preliminary development and evaluation. In Proceedings of the 1990 International Conference on Acoustics, Speech and Signal Processing, April 1990.

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