The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. ex. Some numerals are expressed as "XNUMX".
Copyrights notice
The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. Copyrights notice
본 논문에서는 기계 음역을 위한 새로운 하이브리드 방법을 제안합니다. 우리의 방법은 새로 제안된 1단계 조건부 무작위 필드(CRF) 방법과 잘 알려진 JSCM(Joint Source Channel Model)을 결합한 것을 기반으로 합니다. 본 논문의 기여는 다음과 같다. (2) 기계 음역을 위한 3단계 CRF 모델을 제안한다. 첫 번째 CRF는 입력 단어의 문자열을 덩어리로 분할하고, 두 번째 CRF는 각 덩어리를 대상 언어의 문자로 변환합니다. (4) 1단계 CRF 모델의 공동 최적화 방법과 고속 복호 알고리즘도 제안한다. 우리의 실험에서는 XNUMX단계 CRF 모델의 공동 최적화가 JSCM과 같거나 그보다 더 잘 작동하고 빠른 디코딩 알고리즘이 디코딩 시간을 크게 단축한다는 것을 보여줍니다. (XNUMX) JSCM을 위한 WFST(Weighted Finite State Transducer) 프레임워크 기반의 신속한 개발 방법을 제안한다. (XNUMX) 제안된 XNUMX단계 CRF 모델과 JSCM의 결합은 top-XNUMX 정확도 측면에서 최첨단 결과를 능가합니다.
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
부
Dong YANG, Paul DIXON, Sadaoki FURUI, "A New Hybrid Method for Machine Transliteration" in IEICE TRANSACTIONS on Information,
vol. E93-D, no. 12, pp. 3377-3383, December 2010, doi: 10.1587/transinf.E93.D.3377.
Abstract: This paper proposes a new hybrid method for machine transliteration. Our method is based on combining a newly proposed two-step conditional random field (CRF) method and the well-known joint source channel model (JSCM). The contributions of this paper are as follows: (1) A two-step CRF model for machine transliteration is proposed. The first CRF segments a character string of an input word into chunks and the second one converts each chunk into a character in the target language. (2) A joint optimization method of the two-step CRF model and a fast decoding algorithm are also proposed. Our experiments show that the joint optimization of the two-step CRF model works as well as or even better than the JSCM, and the fast decoding algorithm significantly decreases the decoding time. (3) A rapid development method based on a weighted finite state transducer (WFST) framework for the JSCM is proposed. (4) The combination of the proposed two-step CRF model and JSCM outperforms the state-of-the-art result in terms of top-1 accuracy.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.E93.D.3377/_p
부
@ARTICLE{e93-d_12_3377,
author={Dong YANG, Paul DIXON, Sadaoki FURUI, },
journal={IEICE TRANSACTIONS on Information},
title={A New Hybrid Method for Machine Transliteration},
year={2010},
volume={E93-D},
number={12},
pages={3377-3383},
abstract={This paper proposes a new hybrid method for machine transliteration. Our method is based on combining a newly proposed two-step conditional random field (CRF) method and the well-known joint source channel model (JSCM). The contributions of this paper are as follows: (1) A two-step CRF model for machine transliteration is proposed. The first CRF segments a character string of an input word into chunks and the second one converts each chunk into a character in the target language. (2) A joint optimization method of the two-step CRF model and a fast decoding algorithm are also proposed. Our experiments show that the joint optimization of the two-step CRF model works as well as or even better than the JSCM, and the fast decoding algorithm significantly decreases the decoding time. (3) A rapid development method based on a weighted finite state transducer (WFST) framework for the JSCM is proposed. (4) The combination of the proposed two-step CRF model and JSCM outperforms the state-of-the-art result in terms of top-1 accuracy.},
keywords={},
doi={10.1587/transinf.E93.D.3377},
ISSN={1745-1361},
month={December},}
부
TY - JOUR
TI - A New Hybrid Method for Machine Transliteration
T2 - IEICE TRANSACTIONS on Information
SP - 3377
EP - 3383
AU - Dong YANG
AU - Paul DIXON
AU - Sadaoki FURUI
PY - 2010
DO - 10.1587/transinf.E93.D.3377
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E93-D
IS - 12
JA - IEICE TRANSACTIONS on Information
Y1 - December 2010
AB - This paper proposes a new hybrid method for machine transliteration. Our method is based on combining a newly proposed two-step conditional random field (CRF) method and the well-known joint source channel model (JSCM). The contributions of this paper are as follows: (1) A two-step CRF model for machine transliteration is proposed. The first CRF segments a character string of an input word into chunks and the second one converts each chunk into a character in the target language. (2) A joint optimization method of the two-step CRF model and a fast decoding algorithm are also proposed. Our experiments show that the joint optimization of the two-step CRF model works as well as or even better than the JSCM, and the fast decoding algorithm significantly decreases the decoding time. (3) A rapid development method based on a weighted finite state transducer (WFST) framework for the JSCM is proposed. (4) The combination of the proposed two-step CRF model and JSCM outperforms the state-of-the-art result in terms of top-1 accuracy.
ER -