64 Chengqing Zong, Yujie Zhang, Kazuhide Yamamoto, Masashi Sakamoto and Satoshi Shirai ATR-SLT (Advanced Telcommunications Research Institute Internat

Size: px

Start display at page:

Download "64 Chengqing Zong, Yujie Zhang, Kazuhide Yamamoto, Masashi Sakamoto and Satoshi Shirai ATR-SLT (Advanced Telcommunications Research Institute Internat"

确焱巫
5 years ago
Views:

1 Journal of Chinese Language and Computing 12(1) 语译汉语语写 + * * 2 * * * , 2728 cqzong@nlpr.ia.ac.cn * ATR {yujie.zhang, kazuhide.yamamoto, masashi.sakamoto, satoshi.shirai}@atr.co.jp (utterance) - 1. [18] (robustness) [2,10] G (CRL) e- mail yujie@crl.go.jp

2 64 Chengqing Zong, Yujie Zhang, Kazuhide Yamamoto, Masashi Sakamoto and Satoshi Shirai ATR-SLT (Advanced Telcommunications Research Institute International, Spoken Language Translation Laboratories) Yamamoto Sandglass [12,13] Mckeown (CO-OP)

3 Chinese Utterance Paraphrasing for Spoken Language Translation 65 [7] 1996 Chandrasekar FSG (Finite State Grammar) [3] Dras TAG(Tree-Adjoining Grammars) [4,5] TAG Boguslavsky [1] Sato Kondo [11, 6] Meteer [8], Ramsay [9] [16] 3.

4 66 Chengqing Zong, Yujie Zhang, Kazuhide Yamamoto, Masashi Sakamoto and Satoshi Shirai (1) Ex.1 (2) (pattern) [14] Ex.2 X1/PN /VV /VV X2/VV X3/NN - X1/PN /VV /VV X3/NN X2/VV - X3/NN X1/PN /VV /VV X2/VV (3) (4) Fei Xia The Part-of-Speech Tagging Guidelines for the Penn Chinese Treebank (3.0)

5 Chinese Utterance Paraphrasing for Spoken Language Translation 67 [17] 4. (1) (2) (3) chunk chunk (4) (5) (6) 4.1 (1) % 11.9% 14 (2) # IJ (1) # VA (2) (1) (2) VA

6 68 Chengqing Zong, Yujie Zhang, Kazuhide Yamamoto, Masashi Sakamoto and Satoshi Shirai (3) % 4.2 Chunk % A) B) C) D) E) C) NT NT NP NP (1) (2) (3) 403 (4) chunk (1)(2)(3) chunk

7 Chinese Utterance Paraphrasing for Spoken Language Translation (4) chunk chunk (finite-state transducer, FST) FST 1. FST A B C D (%) A B C A B C 98.9% Chunk chunk chunk chunk (NP) (VP) (PP) [15, 19, 20] PCFG (Probabilistic Context Free Grammar) Chart A α /P P i P(A α i ) = 1 Utter i chunk n NP, VP PP Chart 5 (a) (b) (c)

8 70 Chengqing Zong, Yujie Zhang, Kazuhide Yamamoto, Masashi Sakamoto and Satoshi Shirai (University of Pennsylvania, UPenn) Treebank 244 PCFG Chart 83.1% 89.6% 6.5% 4.4 chunk chunk [15] [19] [15] [19] 9 (Quantity) (Complement) (Sequential Verb, SV) 3 (Pivot word, PW ) 4 (Complement of pivot word, CPW) 1) 2) 3) 4) 1) 4) (1) (2) (3) (4) (5) [17] 3, 4,

9 Chinese Utterance Paraphrasing for Spoken Language Translation n ( n 1 ) 1 (Head) 2 (Body) 2 (Sub-Body) (slot) Head: Type {Interrogative/Declarative/ } : Keywords {Word 1 /Positon, } : Tense {Present/Past/ } : Attribute {Condition/ } Body: Subject; : Adverbial 1 / Adverbial 2 ; : Predicate; : Object1 {Sub-Body}; : Object2; : Quantity; : Complement; : PW; : CPW; : SV 1 / SV 2 ; 2. PW CPW SV 1 SV

10 72 Chengqing Zong, Yujie Zhang, Kazuhide Yamamoto, Masashi Sakamoto and Satoshi Shirai (1) Input: (I-1) { }<T-NP> { }<NP> { }<PP> { }<BA-NP> { }<VP> <BA-NP> <PP> (O-1) (O-2) (O-3) (O-4) (2) : X X X X X (3) VA NP VA NP VA NP VA NP VA NP (4)

11 Chinese Utterance Paraphrasing for Spoken Language Translation 73 (I-1) <Type> = Declarative <Keywords> = { /2 /7 /6} <Tense> = Past; <Attribute> = Null; (O-1) (O-2) (O-3) 5. PCFG FMM (1) 61 57% 46 43% (A) (B) (C) 2

12 74 Chengqing Zong, Yujie Zhang, Kazuhide Yamamoto, Masashi Sakamoto and Satoshi Shirai (%) (2) A, B, C A B C 3. A B C (%) % A B C (%) A B 60.4% 6. HMM

13 Chinese Utterance Paraphrasing for Spoken Language Translation 75 ATR-SLT [1] Boguslavsky, Igor, Nadezhda Frid et. al. Creating a Universal Networking Language Module within an Advanced NLP System. Proc. Coling pp [2] Carroll, John A. Statistical Parsing. In Handbook of Natural Language Processing. Marcel Dekker, Inc [3] Chandrasekar, R., Christine Doran, and B. Srinivas. Motivations and Methods for Text Si,plification. In proceedings of the 16 th COLING pp [4] Dras, Mark. Representing Paraphrasing Using Synchronous TAGs. In proceedings of the 35 th Annual Meeting of the Association for Computational Linguistics (ACL), pp [5] Dras, Mark. A Mets-Level Grammar: Redefining Synchronous TAG for Translation and Paraphrase. In proceedings of the 37 th Annual Meeting of the Association for Computational Linguistics (ACL), pp

14 76 Chengqing Zong, Yujie Zhang, Kazuhide Yamamoto, Masashi Sakamoto and Satoshi Shirai [6] Kondo, Keiko, Satoshi Sato and Manabu Okumura. Paraphrasing by Case Alternation (in Japanese). Information Processing Society of Japan (IPSJ) Journal. Vol. 42, No.3, March pp [7] Mckeown, Kathleen R. Paraphrasing Questions Using Given and New Information. American Journal of Computational Linguistics. Vol. 9, No.1, Jan.-Mar [8] Meteer, Marie, Varda Shaked. Strategies for Effective Paraphrasing. Proc. Coling [9] Ramsay, Allan and Helen Seville. Understanding English Word Order. Proc. Coling pp [10] Samuelsson, Christer, Mats Wiren. Parsing Techniques. In Handbook of Natural Language Processing. Marcel Dekker, Inc [11] Sato, Satoshi. Automatic Paraphrase of Technical Papers Titles (in Japanese). Information Processing Society of Japan (IPSJ) Journal. Vol. 40, No.7, July pp [12] Yamamoto, Kazuhide, Satoshi Shirai, Masashi Sakamoto and Yujie Zhang. S ANDGLASS : Twin Paraphrasing Spoken Language Translation. Proc. ICCPOL 2001, Seoul, Korea. pp [13] Yamamoto, Kazuhide. Machine Translation by Interaction between Paraphraser and Transfer. To appear in Proc. of Coling Aug. 2002, Taipei. [14] Zhang, Yujie, Kazuhide Yamamoto, Chengqing Zong and Masashi Sakamoto. Paraphrasing Utterances by Reordering Words Using Semi-Automatically Acquired Patterns. To appear in proceedings of NLPRS. Nov., 2001, Tokyo, Japan. [15] Zhou, Ming. A Block-Based Robust Dependency Parser for Unrestricted Chinese Text. Preceedings of the Second Chinese Language Processing Workshop. Oct., pp [16] Zong, Chengqing, Hua Wu, Taiyi Huang and Bo Xu. Analysis on Characteristics of Chinese Spoken Language. In Proceedings of the 5 th Natural Language Processing Pacific Rim Symposium (NLPRS). Beijing, China. October, pp [17] Zong, Chengqing, Yujie Zhang, Kazuhide Yamamoto, Masashi Sakamoto and Satoshi Shirai. Approach to Spoken Chinese Paraphrasing Based on Feature Extraction. In Proceedings of the 6 th Natural Language Processing Pacific Rim Symposium (NLPRS). November, Tokyo, Japan. pp [18]...Vol.13, No.2,1999. pp [19]. - J-.., pp [20]..., pp

15 Chinese Utterance Paraphrasing for Spoken Language Translation 77 Chinese Utterance Paraphrasing for Spoken Language Translation Chengqing ZONG + *, Yujie ZHANG* 5, Kazuhide YAMAMOTO*, Masashi SAKAMOTO* and Satoshi SHIRAI* cqzong@nlpr.ia.ac.cn + National Laboratory of Pattern Recognition, Institute of Automation, CAS. P.O.Box 2728, Beijing , China {yujie.zhang, kazuhide.yamamoto, masashi.sakamoto, satoshi.shirai}@atr.co.jp *ATR Spoken Language Translation Research Laboratories Hikaridai Seika-cho Soraku-gun Kyoto , Japan Abstract In a spoken language translation system, when the input utterance can t be successfully parsed and translated, if the system can provide the other possible expressions of the input, it will be very helpful to improve the performance of the translation system. This paper introduces an approach to paraphrasing Chinese utterances. In this approach, an input utterance is first analyzed in terms of phrase structure, dependency of chunks, etc., by using multiple methods. Then, the main features of the input utterance are extracted, and the extraction results are represented by a frame. Finally, other possible expressions of the input are generated based on the analysis results by different methods. Preliminary results are shown in the paper. Keywords: Utterance paraphrasing, Spoken language translation, spoken language parsing 5 Now she is working at the Communications Research Laboratory (CRL) of Japan at the same address as given above, and her new address is: yujie@crl.go.jp.

, CAS-IA CMU Speech Trans, English-Japanese Speech-to-speech translation, S2S Spoken language translation, SLT Spoken dialogue t

, CAS-IA CMU Speech Trans, English-Japanese Speech-to-speech translation, S2S Spoken language translation, SLT Spoken dialogue t ( 11) cqzong@nlpr.ia.ac.cn http://www.nlpr.ia.ac.cn/english/cip/cqzong.htm No.95, Zhongguancun East Road Beijing 100080, China http://www.ia.ac.cn Tel. No.: +86-10-6255 4263 , CAS-IA 2004-5-19 2 11.1 1989