Speech--to--Speech Translation: a Massively Parallel Memory-Based
Approach
describes one of the world's first successful speech--to--speech machine translation systems. This system accepts speaker-independent continuous speech, and produces translations as audio output. Subsequent versions of this machine translation system have been implemented on several massively parallel computers, and these systems have attained translation performance in the milliseconds range.
The success of this project triggered several massively parallel projects, as well as other massively parallel artificial intelligence projects throughout the world.
Dr. Hiroaki Kitano
received the distinguished `Computers and Thought Award' from the International Joint Conferences on Artificial Intelligence in 1993 for his work in this area, and that work is reported in this book.
Reihe
Sprache
Verlagsort
Verlagsgruppe
Illustrationen
Dateigröße
ISBN-13
978-1-4615-2732-9 (9781461527329)
DOI
10.1007/978-1-4615-2732-9
Schweitzer Klassifikation
1 Introduction.- 1.1 Speech-to-Speech Dialogue Translation.- 1.2 Why Spoken Language Translation is So Difficult?.- 1.3 A Brief History of Speech Translation Related Fields.- 2 Current Research Toward Speech-to-Speech Translation.- 2.1 SpeechTrans.- 2.2 SL-TRANS.- 2.3 JANUS.- 2.4 MINDS.- 2.5 Knowledge-Based Machine Translation System.- 2.6 The HMM-LR Method.- 3 Design Philosophy Behind The ? DMDIALOG System.- 3.1 Introduction.- 3.2 Memory-Based Approach to Natural Language Processing.- 3.3 Massively Parallel Computing.- 3.4 Marker-Passing.- 4 The ? DMDIALOG System.- 4.1 Introduction.- 4.2 An Overview of the Model.- 4.3 Speech Input Processing.- 4.4 Memory-Based Parsing.- 4.5 Syntactic/Semantic Parsing.- 4.6 Discourse Processing.- 4.7 Prediction from the Language Model.- 4.8 Cost-based Ambiguity Resolution.- 4.9 Interlingua with Multiple Levels of Abstraction.- 4.10 Generation.- 4.11 Simultaneous Interpretation: Generation while Parsing is in Progress.- 4.12 Related Works.- 4.13 Discussions.- 4.14 Conclusion.- 5 DMSNAP: An Implementation On The Snap Semantic Network Array Processor.- 5.1 Introduction.- 5.2 SNAP Architecture.- 5.3 Philosophy Behind DMSNAP.- 5.4 Implementation of DMSNAP.- 5.5 Linguistic Processing in DMSNAP.- 5.6 Performance.- 5.7 Conclusion.- 6 Astral: An Implementation On The IXM2 Associative Memory Processor.- 6.1 Introduction.- 6.2 The Massively Parallel Associative Processor IXM2.- 6.3 Experimental Implementation I: A Flat Pattern Model.- 6.4 Performance.- 6.5 Memory and Processor Requirements.- 6.6 Enhancement: Hierarchical Memory Network.- 6.7 Experimental Implementation II: Hierarchical Memory Network Model.- 6.8 Performance.- 6.9 Hardware Architecture for Memory-Based Parsing.- 6.10 Conclusion.- 7 Memoir: An Alternative View.- 7.1 Introduction.-7.2 Overall Architecture.- 7.3 Knowledge Sources.- 7.4 Grammatical Inference.- 7.5 Examples Retrieval.- 7.6 Adaptive Translation.- 7.7 Monitor.- 7.8 Preliminary Evaluation.- 7.9 Conclusion.- 8 Conclusion.- 8.1 Summary of Contributions.- 8.2 Future Works.- 8.3 Final Remarks.
