Spoken Language Translation System
Proceedings of the Thirteenth International Joint Conference on Artificial Intelligence. Videos. Pages 1705--1.
© 1993 International Joint Conferences on Artificial Intelligence
Abstract
This videotape introduces a speech-to-speech machine translation system, which translates Japanese speech into English and German speech. Emphasis here is on the speech recognition and automatic translation . The tape also shows an international joint experiment enabling real-time conversations using speech translation systems.
- [Dohsaka, 1990] Kohji Dohsaka. Identifying the Refer- ence of Zero-Pronouns in Japanese Based on Prag- matic Constraint Interpretation. In Proceedings of ECAI-90, 1990.
- [Hasegawa, 1990] Toshiro Hasegawa. Rule Application Control Method in a Lexicon-driven Transfer Model of a Dialogue Translation System. In Proceedings of ECAI-90, 1990.
- [Kikui, 1992] Gen-ichiro Kikui. Feature Structure based Semantic Head Driven Generation. In Proceedings of Coling-92, 1992.
- [Kume et al., 1989] Masako Kume , Gayle K. Sato, and Kei Yoshimoto. A Descriptive Framework for Trans- lating Speaker's Meaning. In Proceedings of European Chapter of ACL-89, 1989.
- [Nagai et al., 1992] Akito Nagai, Jun-ichi Takami, and Shigeki Sagayama. The SSS-LR Continuous Speech Recognition System:lntegrating SSS-Derived Allo- phone Models and A Phoneme-Context-Dependent LR Parser. In Proceedings of ICSLP-92, 1992.
- [Nagata, 1992] Masaaki Nagata. An Empirical Study on Rule Granularity and Unification Interleaving Toward an Efficient Unification-based Parsing System. In Pro- ceedings of Coling-92, 1992.
- [Sagisaka et al., 1992] Yoshinori Sagisaka, Nobuyoshi Kaiki, Naoto lwahashi, and Katsuhiko Mimura. ATR v-talk Speech Synthesis System. In Proceedings of ICSLP-92, 1992.
- [Takami and Sagayama, 1992] Jun-ichi Takami and Shigeki Sagayama. A Successive State Splitting Algorithm for Efficient Allophone Modeling. In Proceedings of ICASSP-92, 1992.
- [Takezawa et al., 1991] Toshiyuki Takezawa, Kenji Kita, J unko Hosaka, and Tsuyoshi Morimoto. Linguistic Constraints for Continuous Speech Recoog- nition in Goal-Directed Dialogues. In Proceedings of ICASSP-91, 1991.
- [Tomabechi, 1991] Hideto Tomabechi. Quasi-Destructive Graph Unification. In Proceedings of ACL-91, 1991.
References were extracted automatically from the PDF and may contain errors.
Full text extracted automatically from the PDF; figures, tables and equations may differ. The PDF is the version of record.
1 Distinctive Features of the System
Our spoken language translation system, called ASURA (Advanced Speech Understanding and Rendering Sys tem of ATR), consists of three basic components: speech recognition, translation and speech synthesis. The video focuses on speech recognition and translation. Speech synthesis is explained in [Sagisaka et al., 1992].
1.1 Speech Recognition
- (1)HMM-LR Speech Recognition is developed and em ployed combining allophonic HMMs (Hidden Malkov Models) and a generalized LR parser extended to handle phoneme context [Nagai et al., 1992] .
- (2)HMMs are generated automatically from statistical data by our Successive State Splitting (SSS) algorithm [Takami and Sagayama, 1992].
- A speaker adaptation technique allows accurate speaker-independent recognition.
1.2 Language Translation
- ( 1) Expressions peculiar to spoken language can be ac cepted and translated properly into the target languages [Kume et al., 1989][Dohsaka, 1990] [Hasegawa, 1990] .
- LingiJistic knowledge sources for translation are de fined declaratively while still maintaining efficiency [N a gata, 1992].
- Declarative grammars are handled efficiently using a new unification algorithm [Tomabechi, 1991] and rule application strategy [Kikui, 1992].
1.3 Integration of Speech Recognition and Language Translation
For efficien'cy, linguistic constraints are distributed throughout speech recognition and language translation
[Takezawa et al., 1991]. Speech recognizer produces sen tential candidates, some of which may be ruled out by the translation component.
@inproceedings{ijcai1993p239,
title = {Spoken Language Translation System},
author = {Kikui, Gen-ichiro and Seligman, Mark and Takezawa, Toshiyuki and Suzuki, Masami and Kita, Kenji and Morimoto, Tsuyoshi and Nagata, Masaaki and Tashiro, Toshihisa and Tropf, Herbert S. and Sagayama, Shigeki and Takami, Jun-ichi and Ohkura, Kazumi and Kurematsu, Akira},
booktitle = {Proceedings of the Thirteenth International Joint Conference on
Artificial Intelligence, {IJCAI-93}},
publisher = {International Joint Conferences on Artificial Intelligence},
editor = {Ruzena Bajcsy},
pages = {1705---1},
year = {1993},
note = {Videos},
url = {https://ijcai.org/proceedings/1993/239},
}
TY - CPAPER TI - Spoken Language Translation System AU - Kikui, Gen-ichiro AU - Seligman, Mark AU - Takezawa, Toshiyuki AU - Suzuki, Masami AU - Kita, Kenji AU - Morimoto, Tsuyoshi AU - Nagata, Masaaki AU - Tashiro, Toshihisa AU - Tropf, Herbert S. AU - Sagayama, Shigeki AU - Takami, Jun-ichi AU - Ohkura, Kazumi AU - Kurematsu, Akira T2 - Proceedings of the Thirteenth International Joint Conference on Artificial Intelligence, IJCAI-93 PB - International Joint Conferences on Artificial Intelligence PY - 1993 SP - 1705 EP - -1 UR - https://ijcai.org/proceedings/1993/239 ER -Copied
Machine-readable versions of this paper, for tools and AI agents:
