Case Study

Post-Edited Neural Machine Translation applied

How NMTPE can help solving common business challenges

A applied sample of neural translation memory and post -editing applied

Our Numbers Say It All

Executive Summary

Avionics client purchase planes from a foreign government to perform in-flight training with local teams.

The documentation, technical manuals and any other related documentation of the planes was only available in hard copy foreign language documents.

 

Word s translated

Month time

Hard copy Pages

Challenge

  • Hard copy documentation in French.
  • Nearly 38000 pages of content including extensive technical drawings.
  • 6 months timeline for FAA approvals.
  • Very Limited Budget.

A traditional processing  would have required 12-16 months processing and respective cost associated with 6.8 Million words of translation.

How our NMT PE Solution helped

We designed a process to enable Post Edited Machine Translation rather than traditional processing, based on the clients scope of project.

  1. Scan and OCR Hard Copy Documentation
  2. Generate Digital and machine readable versions of original Documentation
  3. Build initial NMT training repository based on materials and reference documentation
  4. Create and Train Neural Machine translation engine
  5. Process files through Post edited Neural Machine translations
  6. Process Layout Adjustments NMTPE Documents  on 38000 Pages
  7. Deliver all material within 6 months.

Results

  • Delivery of Final NMTPE processed documents within 5.5 Months
  • Cost level to client at 25% of traditional translation cost

E

“Lorem ipsum dolor sit amet, consectetuer adipiscing elit. Donec odio. Quisque volutpat mattis eros.”

rolf hecken

Are you ready to partner with us?

Lorem ipsum dolor sit amet, consectetuer adipiscing elit.

Other Technology Highlights

Lorem ipsum dolor sit amet, consectetuer adipiscing elit.