If you decide to implement machine translation, the first step is to select a machine translation engine.
If you make the wrong choice, even if you invest in the implementation, it could lead to the unfortunate result of the output being unusable.
There are over 40 types of machine translation engines available in the market.
Among the many options, how should we choose the most suitable engine?
This time, I will share the key points.
First, when choosing a machine translation engine,
- Type of Engine (Rule-Based / Statistical / Hybrid)
- Implementation Costs, Operating Costs
- Integration with translation support tools (CAT tools) such as Trados
- Supported File Formats
- Support System
- Usability
It is necessary to confirm points like these.
And what is most important is the translation quality of the machine translation engine.
The quality varies greatly depending on the engine.
For reference, let's take a look at the results comparing the quality of three engines in German.
The blue part of the graph represents the proportion of text that is "Excellent/Good",
The gray part represents the proportion of text that is "Medium/Poor".
For more information on the quality evaluation criteria for machine translation, please see here.
In Engine B, nearly half of the texts were of "poor quality," but,
in Engine C, almost all of the texts became "good quality."
Even if the same text is translated in this way, there can be significant differences in quality depending on the machine translation engine used.
Therefore, we recommend conducting sample translations with multiple engines in advance and making a decision after comparing the quality.
At Human Science, we also provide consulting for the selection of machine translation engines.
We will conduct quality evaluations on 3 to 5 candidate engines selected from multiple engines,
calculate cost reduction rates, translation duration reduction rates, return on investment, and more after implementation,
and propose the optimal engine.
Please feel free to contact us.
If the form is not available, please send your inquiry via email to hsweb_inquiry@science.co.jp.
Alternatively, please feel free to contact us by phone at TEL: 03-5321-3111.