| 
   
Article Details
Global Watchtower
Common Sense Advisory Blogs
Building the Localization Web
Posted by Donald A. DePalma on December 5, 2013  in the following blogs: Technology
Pages | |


Smaller language service providers (LSPs) process fewer words than larger ones. Not only does this mean that they earn less income than their bigger competitors, but it also puts them at a disadvantage when it comes to leveraging linguistic assets due to the smaller size of their terminology databases and translation memories (TMs). These less comprehensive language resources limit reuse on subsequent projects or for training statistical machine translation (SMT) software. 

It was this argument that won the Falcon consortium of European language technology developers and their academic partners a €1.5 million grant from the European Commission’s Seventh Framework Project (FP7) for a two-year project that began in October. Their project will create an open web platform that allows LSPs to achieve greater leverage by combining their termbases and TMs from their projects with those of public bodies such as the European Parliament, their clients, and even other LSPs. The winning consortium includes Easyling, Interverbum Technology, XTM International, and Ireland’s CNGL: Centre for Global Intelligent Content (formerly known as the “Centre for Next-Generation Localisation”).

  • Step 1: The platform. Falcon’s first deliverable will let users search multiple termbases and TMs through a single interface. Individual URLs will identify terms and segments, and their metadata. What enables this is a data model that “federates” these disparate linguistic assets into what appears to the user as a single web-based database. This work is based on existing specifications from the W3C, including its Semantic Web, Resource Description Framework (RDF), and the Simple Protocol and RDF Query Language (SPARQL). Access will be control based on rules developed by a CNGL academic partner, Trinity College Dublin

    What all this means is that users can access disparate linguistic assets wherever they are stored, once the owners tag them for use within this framework. This virtual “play it where it lays” approach is a more achievable alternative to strategies that consolidate multiple termbases and TMs into a single centralized database. Such massive centralization might work at large corporate users when mandated and adequately funded, but few LSPs would be able to undertake such an effort. The Falcon approach will allow an LSP to create an on-demand database of assets within and among its clients. 

  • Step 2: Training. Once its federated platform is in place, the Falcon team will pursue a more ambitious agenda, including: 1) on-demand training of language technologies such as machine translation based on this federated model; and 2) the potential for machine-human interaction during QA and review cycles with the goal of developing a continuous self-training model for those assets. 

    Success in delivering the training components will benefit the many LSPs that lack the budget or staffing to increase their use of more sophisticated translation automation technologies (see “Tech-Savvy Language Service Providers,” Aug10). 

LSPs and buyers both stand to benefit from the FP7 funding for Falcon and its sister project, Lider, which will develop use cases and guidelines for best practices. Both projects support the European Union’s Lisbon Strategy to make Europe the “most dynamic competitive knowledge-based economy in the world.” Investment in projects such as Falcon will enable the sharing of knowledge across the multilingual bloc, but many entrepreneurs will also need the involvement of the private sector to fully develop and productize these essential technologies.



 

Post a Comment

Name
Email address :(Your Email Address Will Not Be Displayed)
URL

Your Comments
Enter Code given below :    

Related Research
Tech-Savvy Language Service Providers
The Case for Terminology Management
Link To This Page

Bookmark this page using the following link:http://www.commonsenseadvisory.com/Default.aspx?Contenttype=ArticleDetAD&tabID=63&Aid=5598&moduleId=390

Do you have a website? You can place a link to this page by copying and pasting the code below.
Back
Keywords: Machine translation, Terminology management, Translation, Translation technologies

  
Refine Your Search
  Date
Skip Navigation Links.
Skip Navigation Links.




 
 
Terms of Use | Privacy Statement | Contact Us
Copyright © 2014 Common Sense Advisory, Inc. All Rights Reserved.