DeMT™ Evaluate
This is not your typical MT evaluation. We don’t help you choose the best out-of-the-box MT, we show you the potential of MT customization with excellent data.
Measure the impact of your customized machine translation models
Engine customization is a well-known method for increasing the quality of MT output. It is also a way to ensure cost efficiency and optimal use of your MT technology. When embarking on the customization journey, you want to have an established way of evaluating and benchmarking the customized MT output and assessing the potential ROI. We make that process easy.
Actionable insights into your MT performance
We combine advanced automated metrics and human evaluation to quantify the impact of training and customization.
Evaluation workflow
Achieve measurable MT quality improvements by adding periodic, streamlined evaluation workflow to your MT pipeline. We can help you identify the areas of refinement and monitor the progress over time.
Gold standard test sets
Multifaceted engine evaluation
Detailed evaluation report
Training dataset creation
Get a glimpse of how we do it
To prove that customization is your best bet and not just our bold claim, our reports include BLEU and COMET scores and the relative and absolute improvements compared to the generic MT output.
Filter by:
Language pair
Financial Services
Pharmaceuticals & Biotechnology
Retail & Wholesale Distribution/E-Commerce
BLEU Scores
TAUS DeMT™️ is not limited to the domains or language pairs shown above. Want to add your language pair and domain?
Contact us
DeMT™ Evaluation Report, June 2022