Rapidly develop KantanMT engines that are production ready
One of the biggest challenges in customizing Statistical Machine Translation systems is rapidly improving the engine after its initial training.
While for the most part, you can build a baseline engine using existing Translation Memory - the real challenge is how do you go beyond this and achieve higher levels of quality.
More importantly, how can you do this rapidly and with minimum cost and effort?
Only Kantan BuildAnalytics provides functionality to manage:
- Fluency Analysis: Work with detailed segment level BLEU scores to determine how relevant your training data is and how it impacts KantanMT engine fluency
- Recall & Precision Analysis: Analyse segment level F-Measure scores to determine the recall precision of your KantanMT engines
- Estimate Post-Editing: Determine post-editing effort by analysing segment level TER scores
- Gap Analysis: Detect data gaps within a KantanMT engine and take corrective action quickly using our Gap Analysis technology
- Training Reject Reports: Work with detailed training data analysis to determine the suitability and relevancy of training data
- Timeline & Version Control: Work with a chronological timeline of activities for each engine and use version control for precise management of released and production-ready engines.
Here's what our members are saying:
I can't think of any way it could be any simpler.
Ilan Bloch, Director, Baguette Translations
We ran MT tests into 12 languages and the results were extraordinary.
Marcello Rizzo, Architect, Medialocate