依存结构作为语义分析的先决工作，有助于提高统计机器翻译的质量，为了解决统计机器翻译中的这一关键技术，拓展藏文信息处理的新研究领域，更好地为国内藏语机器翻译需求做服务。本项目在已经构建的大规模规范藏语语料库及现有研究基础上，将研究藏语依存句法树分析和处理技术，提出符合藏语句法特征的翻译模型和翻译方法，最终实现基于依存树到串藏语句法统计机器翻译系统。拟解决以下四个方面的问题：一是藏语依存句法标注规范的确立与藏语句法树库的构建技术；二是改进和完善项目组现有的藏语依存树分析工具；三是藏语开放词和封闭词的划分，通过研究头依存关系的抽取技术，设计并实现藏语句法翻译模型的训练工具；四是设计依存结构树到串翻译模型的解码算法,并实现翻译模型的解码器。

The dependency structure as a first step towards semantics, is believed to be helpful to improve translation quality. So in order to solve these key technologies of the Tibetan syntacitc statistical machine translation(SMT), to develop new research area of the Tibetan language information processing, and to better serve demand for the domestic Tibetan machine translation. We will analyse the Tibetan dependency syntax tree and its processing technology on normal large-scale corpus of the Tibetan and Chinese, then given translation model and decoding methods for Tibetan syntactic features, finally we will implement the Tibetan to Chinese SMT system, which used dependency to string model. For this reason, the project group intends to handle the problems of the following four aspects. Firstly, we will completely finish the definition of Tibetan dependency syntax grammars. Secondly, to improve and perfected our previous Tibetan dependency parser. Thirdly,to divide Tibetan word tags to two classifications of opening and closing word, study the technology of how to extract the head relation rules in dependency tree, and then we will develop the training tool for translation model. Fourth, we will design the decoding algorithm that used to decode dependency structure tree-to-string translation model, and after to implement syntactic SMT model decoder.