Media.Player

Customized Solutions for Domain Oriented Hybrid Machine Translation

总结

基于新一代混合机器翻译技术,面向领域提供机器翻译定制化解决方案。
Beijing CCID TransTech Co., Ltd. (hereinafter referred to as “CCID TransTech”) is affiliated to China Center for Information Industry Development (CCID). CCID TransTech can trace its history back to the Chinese Information Processing Open Lab (CIPOL), which was founded by Chen Liwei, an Academician of the Chinese Academy of Engineering in 1987.
CCID CIPOL
An Institute combined R&D of MT/CAT with translation services.
i.e. MT R&D + Applications + MT-based Services 采用国际先进的融合统计、规则与模版的新一代混合机器翻译架构设计
 采用独创的基于大规模语言知识库的机器翻译专利技术
 采用基于Web词库和可比语料库的智能数据挖掘技术
 自动翻译功能：适用于常年大量阅览外文资料的科研部门
 辅助翻译平台：适用于需要精确翻译的项目团队或个人
 双语资源管理平台：适用于有大量专业术语及双语资料整理需求的科研机构和管理机构
 定制化系统翻译准确率可达80%以上

Strength of CCID CIPOL
1) Background: CCID has to Strengthen the R&D of MT/CAT Tools
CCID needs to translate a large amount of research reports and user manuals on ICT fields each year. The documents include millions of words to be translated from Chinese/English into English/Chinese every year.

Given the stringent requirement on translation quality and the tight schedule and most of the low-quality market products cannot meet our strict criteria, CCID has to build bilingual corpora and develop its own high-quality MT systems.


2）Undertaking large-scale R&D of machine translation project
In the past years, CCID CIPOL had established collaborative relations with big organization such as Beijing Organizing Committee for the Games of the XXIX Olympiad (BOCOG), the executive body of EU-China Information Society Project and many others.

In most of the projects, a large amount of information and working documents were translated at the initial stage, through pre-processing of the MT tools, to extract sentence templates and to ensure terminology consistency.

CCID CIPOL's Resource Advantages
1. CCID CIPOL accumulates a contemporary Chinese-English bilingual parallel corpora in the field of ICT with more than one million sentence pairs each year.

2. CCID CIPOL has accumulated 80,000 translation templates and 16 million entries used for Hybrid Machine Translation.

3. CCID has been exploiting and using Multi-word expressions (MWEs) and translation templates to improve machine translation performance.

4. The data resource for research is not confined to the academic interests, but with a strong orientation in application.（1）采用当代国际先进的混合机器翻译结构；
（2）采用当代面向产业应用的以规则系统为骨架的混合机器翻译结构设计；
（3）采用数据驱动方法从语料库提取规则与模板的技术；
（4）采用数据驱动方法从语料库提取多词表达的技术。（1）全自动高质量宽覆盖面的机器翻译研发还需要相当长的时间；
（2）实用化、高效率的数据驱动与知识驱动的融合技术仍在研究之中。

详细技术说明

（1）通过定制化开发可以满足特定领域用户使用需求；
（2）正在改进和发展以规则系统为主体融合统计方法的混合机器翻译系统；
（3）根据领域不同用户需求不同，定制化研发周期为3个月到1年不等。

预期达到的市场价值 Prospective market value/business value
随着全球化和互联网迅速发展，跨语言的网络资源呈几何级数增长，迅速改变着信息传播的方式，极大地刺激了机器翻译产业的发展。由于目前在线统计机器翻译普遍存在固有的技术缺陷和数据安全隐患问题，从而为本项目定制化开发的产品应用与推广带来了极大的空间。同时，随着本项目翻译引擎技术的不断改进和发展，也将促进机器翻译与信息检索、信息抽取以及企业生态系统等融合的无缝集成应用的发展。鉴于以上所述，本项目具有广阔的应用前景以及极其可观的市场价值。

合作类型

技术售让

覆盖范围

信息技术、通讯、数字技术

商品和服务

其他