Small: RI: Broad-Coverage High-Accuracy Machine Translation into Morphologically-Rich Languages

小：RI：广泛覆盖的高精度机器翻译为形态丰富的语言

• 批准号: 0915327
• 负责人: Alon Lavie
• 金额: $ 45万
• 依托单位: Carnegie-Mellon University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2009
• 资助国家: 美国
• 起止时间: 2009-09-01 至 2013-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=0915327&HistoricalAwards=false
• 关键词: Small; RI; Broad; Coverage; Accuracy

Machine Translation (MT) into morphologically-rich languages poses uniquechallenges that have so far not been adequately addressed in state-of-the-artapproaches. Even the best available MT systems into languages such as Arabicfrequently produce translations that are disfluent and lack proper grammaticalstructure. This project explores novel approaches that address these issues bythe development of a statistical MT framework that incorporates deeper levelsof modeling of syntax and morphology. While the methods explored are largelylanguage independent, the research is conducted and experimentally evaluatedwithin the context of a large-scale English-to-Arabic MT system constructedusing vast corpora available from LDC.The research in this project focuses on novel approaches for combiningsyntactic and non-syntactic translation resources that are automaticallyacquired from vast amounts of parallel data and on exploring severalalternative pathways for the integration of information provided by ahigh-accuracy morphological analysis and generation engine for Arabic into theMT framework. The project also explores methods for improving the syntax of MToutput in Arabic using syntactic transfer rules that model syntacticdivergences between English and Arabic. The goal is to develop anEnglish-to-Arabic MT system that produces significantly more fluent,grammatical and accurate Arabic output than the current best systems, asmeasured by MT evaluation metrics (such as BLEU and METEOR), and as judged byhuman evaluators.The availability of high-accuracy fully-automatic Machine Translation fromEnglish into Arabic has high potential value to the Arabic-speaking populationat large, by opening up access to all English content available over the web.Such high-quality MT into Arabic may potentially also improve access tomarkets in the Arabic-speaking world for US and international companies.

机器翻译（MT）为形态丰富的语言带来了独特的挑战，迄今为止，最先进的方法尚未充分解决这些挑战。 即使是最好的机器翻译系统，例如阿拉伯语，也经常产生不流畅且缺乏正确语法结构的翻译。 该项目通过开发统计机器翻译框架来探索解决这些问题的新方法，该框架结合了更深层次的语法和词法建模。 虽然所探索的方法在很大程度上与语言无关，但该研究是在使用 LDC 提供的大量语料库构建的大规模英语到阿拉伯语 MT 系统的背景下进行和实验评估的。该项目的研究重点是结合从大量并行数据中自动获取的句法和非句法翻译资源的新方法，并探索几种集成翻译资源的替代途径。 由高精度阿拉伯语形态分析和生成引擎提供的信息进入 MT 框架。该项目还探索了使用模拟英语和阿拉伯语之间句法差异的句法转移规则来改进阿拉伯语 MToutput 语法的方法。我们的目标是开发一个英语到阿拉伯语的机器翻译系统，根据机器翻译评估指标（例如 BLEU 和 METEOR）以及人类评估者的判断，该系统能够比当前最好的系统产生更加流畅、语法和准确的阿拉伯语输出。通过开放，从英语到阿拉伯语的高精度全自动机器翻译的可用性对广大阿拉伯语人群具有很高的潜在价值。 访问网络上提供的所有英语内容。这种高质量的阿拉伯语机器翻译也可能会改善美国和国际公司进入阿拉伯语世界市场的机会。