Generating Chinese Couplets using a Statistical MT Approach

6594

Authors

nan

Abstract

Part of the unique cultural heritage of China is the game of Chinese couplets (duìlián). One person challenges the oth- er person with a sentence (first sentence). The other person then replies with a sen- tence (second sentence) equal in length and word segmentation, in a way that corresponding words in the two sentences match each other by obeying certain con- straints on semantic, syntactic, and lexi- cal relatedness. This task is viewed as a difficult problem in AI and has not been explored in the research community. In this paper, we regard this task as a kind of machine translation process. We present a phrase-based SMT approach to generate the second sentence. First, the system takes as input the first sentence, and generates as output an N-best list of proposed second sentences, using a phrase-based SMT decoder. Then, a set of filters is used to remove candidates vi- olating linguistic constraints. Finally, a Ranking SVM is applied to rerank the candidates. A comprehensive evaluation, using both human judgments and BLEU scores, has been conducted, and the re- sults demonstrate that this approach is very successful. 1

Keywords

nan

Full Content

Proceedings of the 22nd International Conference on Computational Linguistics (Coling 2008), pages 377–384 Manchester, August 2008 Generating Chinese Couplets using a Statistical MT Approach Long Jiang Microsoft Research Asia Sigma Center, No. 49, Zhichun Road Haidian District, Beijing 100190, PRC longj@microsoft.com Ming Zhou Microsoft Research Asia Sigma Center, No. 49, Zhichun Road Haidian District, Beijing 100190, PRC mingzhou@microsoft.com Abstract Part of the unique cultural heritage of China is the game of Chinese couplets (duìlián). One person challenges the oth- er person with a sentence (first sentence). The other person then replies with a sen- tence (second sentence) equal in length and word segmentation, in a way that corresponding words in the two sentences match each other by obeying certain con- straints on semantic, syntactic, and lexi- cal relatedness. This task is viewed as a difficult problem in AI and has not been explored in the research community. In this paper, we regard this task as a kind of machine translation process. We present a phrase-based SMT approach to generate the second sentence. First, the system takes as input the first sentence, and generates as output an N-best list of proposed second sentences, using a phrase-based SMT decoder. Then, a set of filters is used to remove candidates vi- olating linguistic constraints. Finally, a Ranking SVM is applied to rerank the candidates. A comprehensive evaluation, using both human judgments and BLEU scores, has been conducted, and the re- sults demonstrate that this approach is very successful. 1 Introduction Chinese antithetical couplets, called “duìlián”, form a special type of poetry composed of two © 2008. Licensed under the Creative Commons Attri- bution-Noncommercial-Share Alike 3.0 Unported license (http://creativecommons.org/licenses/by-nc- sa/3.0/). Some rights reserved. sentences. They use condensed language, but have deep and sometimes ambivalent meanings. The two sentences making up the couplet are called the “first sentence” (FS) and the “second sentence” (SS) respectively. Chinese couplets are considered an important cultural heritage. A couplet is often written in calligraphy on vertical red banners, and typically placed on either side of a door or in a large hall during special occasions such as wedding cere- monies and the Chinese New Year. People also use couplets to celebrate birthdays, mark the openings of a business, and commemorate histor- ical events. Chinese couplets have also been used effectively in teaching Chinese in China. An example of a Chinese couplet is “海 阔 凭 鱼 跃; 天 高 任 鸟 飞”, where the FS is “海 阔 凭 鱼 跃” and the SS is “天 高 任 鸟 飞”. It says that the sea is wide enough so that fish can jump at their pleasure, and the sky is high enough so that bird can fly unrestrictedly. The correspondence between individual words of the FS and SS is shown here: 海 阔 凭 鱼 跃 sea wide allow fish jump | | | | | 天 高 任 鸟 飞 sky high permit bird fly Figure 1. An Example of a Chinese Couplet. Generating the SS of a …

Content truncated to 500 words

Labeled Text

Part of the unique cultural heritage of China is the game of Chinese couplets (duìlián). One person challenges the other person with a sentence (first sentence). The other person then replies with a sentence (second sentence) equal in length and word segmentation, in a way that corresponding words in the two sentences match each other by obeying certain constraints on semantic, syntactic, and lexical relatedness. This task is viewed as a difficult problem in AI and has not been explored in the research community. In this paper, we regard this task as a kind of machine translation process. We present a phrase-based SMT approach to generate the second sentence. First, the system takes as input the first sentence, and generates as output an N-best list of proposed second sentences, using a phrase-based SMT decoder. Then, a set of filters is used to remove candidates violating linguistic constraints. Finally, a Ranking SVM is applied to rerank the candidates. A comprehensive evaluation, using both human judgments and BLEU scores, has been conducted, and the results demonstrate that this approach is very successful. 1

Publication Details

Paper ID:
6594
Year:
2008
Publication:
TIPSTER
Publication Type:
conf
DOI:
nan