Phonology and Morphology Modeling in a Very Large Vocabulary Hungarian Dictation System(Speech and Hearing)
スポンサーリンク
概要
- 論文の詳細を見る
This article introduces a novel approach to model phonology and morphosyntax in morpheme unit-based speech recognizers. The proposed methods are evaluated on a Hungarian newspaper dictation task that requires modeling over 1 million different word forms. The architecture of the recognition system is based on the weighted finite-state transducer (WFST) paradigm. The vocabulary units used in the system are morpheme-based in order to provide sufficient coverage of the large number of word-forms resulting from affixation and compounding. Besides the basic pronunciation model and the morpheme yV-gram language model we evaluate a novel phonology model and the novel stochastic morphosyntactic language model (SMLM). Thanks to the flexible transducer-based architecture of the system, these new components are integrated seamlessly with the basic modules with no need to modify the decoder itself. We compare the phoneme, morpheme, and word error-rates as well as the sizes of the recognition networks in two configurations. In one configuration we use only the TV-gram model while in the other we use the combined model. The proposed stochastic morphosyntactic language model decreases the morpheme error rate by between 1.7 and 7.2% relatively when compared to the baseline trigram system. The proposed phonology model reduced the error rate by 8.32%. The morpheme error-rate of the best configuration is 18% and the best word error-rate is 22.3%.
- 社団法人電子情報通信学会の論文
- 2004-12-01
著者
-
Furui Sadaoki
Department Of Computer Science Graduate School Of Information Science And Engineering Tokyo Institut
-
SZARVAS Mate
Department of Computer Science, Tokyo Institute of Technology
-
Szarvas M
Department Of Computer Science Tokyo Institute Of Technology
関連論文
- Gait-based Person Identification Robust against Speed Variation using CHLAC features and HMMs
- Gait-based Person Identification Robust against Speed Variation using CHLAC features and HMMs
- Gait-based Person Identification Robust against Speed Variation using CHLAC features and HMMs
- Tree-Structured Clustering Methods for Piecewise Linear-Transformation-Based Noise Adaptation(Speech and Hearing)
- Robust Scene Extraction Using Multi-Stream HMMs for Baseball Broadcast(Image Processing and Video Processing)
- Automatic recognition of Indonesian declarative questions and statements using polynomial coefficients of the pitch contours
- Initial evaluation of the drivers' Japanese speech corpus in a car environment (Speech) -- (国際ワークショップ"Asian workshop on speech science and technology")
- Accent analysis for Mandarin large vocabulary continuous speech recognition (Speech) -- (国際ワークショップ"Asian workshop on speech science and technology")
- Evaluation of a Noise-Robust Multi-Stream Speaker Verification Method Using F_0 Information
- Topic Extraction Based on Continuous Speech Recognition in Broadcast News Speech
- Noise Robust Speech Recognition Using F_0 Contour Information(Speech Dynamics by Ear, Eye, Mouth and Machine)
- Phonology and Morphology Modeling in a Very Large Vocabulary Hungarian Dictation System(Speech and Hearing)
- 連続発話認識のための言語モデル
- Dynamic Bayesian Network-Based Acoustic Models Incorporating Speaking Rate Effects(Speech and Hearing)
- Neural-network-based HMM adaptation for noisy speech recognition
- Speaker Verification Using MMAP Adaptation (言語理解とコミュニケーション)
- Speaker Verification Using MMAP Adaptation (音声)
- Subject Adaptation and Adaptive Training for Gait-based Person Identification
- Subject Adaptation and Adaptive Training for Gait-based Person Identification
- Two-pass Approach for Recognizing Code-Switching Speech
- Two-pass Approach for Recognizing Code-Switching Speech
- Two-pass Approach for Recognizing Code-Switching Speech
- Subject Adaptation and Adaptive Training for Gait-based Person Identification
- Speaker Verification Using MMAP Adaptation
- Two-pass Approach for Recognizing Code-Switching Speech