Multi-engine machine translation
Machine translation with multiple systems
Multi-engine machine translation is an approach for evaluating and picking a machine translation engine. With multi-engine machine translation, the same input is sent to multiple machine-translation systems. The translations are then ranked to select the best one.
The challenges of multi-engine machine translation are related to the translation output and the implementation:
- Lack of consistency across segments
- Increased complexity, latency, and cost
The goal of multi-engine machine translation is to get higher quality translations than any single engine could achieve.