TAUS/CNGL Machine Translation Post-Editing (MTPE) Guidelines published - Pangeanic participated therein

TAUS and CNGL Release Baseline MTPE Guidelines

TAUS made available some baseline Machine Translation Post-Editing (MTPE) Guidelines on their website on 3rd January 2011. Produced in partnership with the CNGL (Centre for Next Generation Localisation) in Dublin, the document results from an industry-wide initiative whose core working group consisted of:

• Fred Hollowood (Symantec, and TAUS Advisory Board)
• Sharon O'Brien (Dublin City University)
• Rahzeb Choudhury (TAUS Coordinator)
• Jaap van der Meer (TAUS Supervisor)

We are now happy to help in disseminating these Guidelines—which are also available as a downloadable PDF file—since they will surely shed some much-needed light on a very hot topic in our industry.

Pangeanic’s Active Role and Expertise

Pangeanic is proud to have directly contributed to this standard. Elia Yuste (Pangeanic) was one of the industry experts who took part in the special discussion meeting held at the TAUS User Conference in Portland last October. She also provided deep operational feedback in the consulting round that followed in November 2010 to systematically review the draft guidelines.

For more information on the full list of participants and the step-by-step collaborative process for arriving at these guidelines, please check out the official TAUS repository.

Redefining Translation Quality Thresholds

These guidelines are not meant to rigidly address every single MTPE scenario, but rather to help interested organizations figure out, benchmark, and adapt MTPE procedures that really work for their specific business needs.

Apart from offering general recommendations to minimize total post-editing effort, the document makes an interesting and vital distinction between two quality tiers:

1. Guidelines aimed at reaching a cost-effective "good-enough quality" for rapid informational access.
2. Guidelines designed to achieve a final quality that is equal or similar to that produced by a human translator from scratch.

Pangeanic’s MTPE and PangeaMT Programs

For your information, Pangeanic runs its own dedicated MTPE Programme. We approach this from two distinct perspectives:

• As an LSP: Leveraging long-term, practical experience in post-editing raw output from different types of machine translation systems.
• As a Technology Developer: Through the lens of PangeaMT, Pangeanic's Customized Machine Translation solutions division.

We will not only create your tailor-made, in-domain PangeaMT engine, but we will also take care of your ongoing MTPE needs. Furthermore, if you are another LSP who has commissioned PangeaMT engines from us to speed up your specialized translation projects, we will be delighted to provide you with expert MTPE orientation and training frameworks.