KantanLQR is a feature of the KantanMT™ platform that is used to carry out human evaluation of Machine Translation (MT) output. KantanLQR was introduced with the main purpose of speeding up the evaluation process by automating the steps involved, especially the analysis of the results.
With KantanLQR, you can perform two kinds of evaluation: comparative and qualitative. Comparative evaluation (namely, A/B Testing) is used to compare up to four different outputs and rank them according to their overall quality. Quality evaluation is used to score segments based on preferred Key Performance Indicators (KPIs) which can be pre-selected by the user. Machine Translation segments can also be assessed along side a specified image.
You can access KantanLQR by logging onto KantanMT. On the left-hand side menu, under KantanLQR there are three options: Dashboard, KPIs, and Reviewers.
In the KantanLQR Dashboard, you can see your projects, create new projects (if you are a project manager), review projects (if you are a reviewer) and check the progression of your projects. In the KPIs and Reviewers tabs, project managers can check respectively which KPIs and which reviewers are employed for their account.
For more information on how to create or review KantanLQR projects, check the following how-to guides:
- How to Create a New Quality Evaluation Project in KantanLQR™
- How to Create a New A/B Test Project in KantanLQR™
- How to Review a New AB Test Project in KantanLQR™
LQR for project managers:
LQR for reviewers: