Testing of translation solutions for industrial applications

Service for the evaluation of text-to-text and speech-to-text translation systems, including systems for automatic subtitling, used for industrial applications (e.g. manuals’ translation, HCI, robot interactions, ….). Multi-faceted evaluations relying on different types of benchmarks are applied to assess system behavior on different aspects (e.g. terminology or gender translation), to compare the performance of different solutions, or track the performance of a system over time. Two main types of evaluation are supported: – Automatic evaluation, through a pool of standard metrics able to capture different definitions of translation quality. – Human evaluation, through highly configurable tasks relating quality assessment to the specific needs of the users of the evaluated solutions