In mid-November 2016, Google Neural Machine Translation (GNMT) became available for Japanese.
The translation accuracy has significantly improved compared to the previous Google Translate for both English-Japanese and Japanese-English language pairs.
"It is said that machine translation has improved in accuracy, but what is the reality?"
This is a question many readers have.
Unique Research Results of Human Science
We immediately began investigating GNMT.
We examined the issues and their solutions, and started providing GNMT implementation solutions by the end of the same month.
Our consultants and engineers listen to your needs and provide implementation and customization services for customers who want to utilize GNMT in-house.
We have achieved a 34% cost reduction and a 25% reduction in duration in actual projects using GNMT.
No technology can be a mature service from the start.
When providing new services to customers, verification is an important process.
In this article, we have compiled examples and issues in practical translation by category: IT, medical, and mechanical. (The GNMT translation is based on the time of writing this article.)
In the upcoming serialized articles, we will share the verification results.
Related Articles
Human Science Machine Translation Solutions