In the previous topic, we covered the
machine translation using neural networks which is currently receiving a lot of attention.
Even free translation services such as Google Translate and Microsoft Translator
are adopting neural network mechanisms and becoming a hot topic.

By the way, everyone, have you ever wondered how much the translation accuracy will improve by adopting neural networks?

So, in this topic, we will introduce the results of comparing the accuracy of traditional statistical-based machine translation engines and neural network-based machine translation engines.

 

●Accuracy Comparison Results

Machine translation is here in Japanese.

——————————————–
This command will be applied to the displayed content after the command is executed, so if an error message is displayed, it will be applied to that error message.
——————————————–

Let's compare the results of translating this Japanese using a corporate " statistical-based MT engine " and a " neural network-based MT engine " into English.
(Both use the same corporate MT engine.)

A long sentence that at first glance does not seem suitable for machine translation,
but what kind of result will come out?

・Machine translation results from a statistical-based MT engine

——————————————–
This command applies to the contents of the after running the command,
error message is displayed if the applies to the error message.
——————————————–

・Neural Network MT Engine Machine Translation Results

——————————————–
This command applies to the displayed content
after the command execution, so if you receive an error message,
it will be applied to the error message.
——————————————–

How about comparing accuracy?

 

●Consideration of Accuracy Comparison Results

In the machine translation results of the statistical-based MT engine,
“the contents of the after running the command”,
“if the applies to the error message”,
there are grammatical errors and the sentences do not make sense.

On the other hand, in the machine translation results of the neural network MT engine,
there may be some unnatural parts,
but compared to the statistical-based MT engine, the grammar is correct and
you can see that the meaning in Japanese is conveyed.

In this way, the MT engine that adopts the mechanism of neural network,
compared to the statistical-based MT engine,
has greatly improved output accuracy!!

 

●Why is a neural network amazing?

In many traditional MT engines, sentences are divided into individual words and translated, then combined using a mechanism called "phrase-based".

On the other hand, in the new engine that adopts the mechanism of neural networks,
instead of simply replacing words with translations,
it is possible to consider the context based on the connections with other words in the same sentence,
and determine how each word should be translated.
This is the idea behind it.

Thanks to this mechanism, we are now able to find more accurate translation candidates than the conventional version, and as a result, we can achieve a quality that is closer to human translation.

If this level of high quality can be achieved,
it seems that the introduction of machine translation will continue to advance even in Japan.

Why not consider introducing machine translation for translating items that require shortening delivery time and reducing costs?
At Human Science, we also introduce securely usable
neural network machine translation engines.

Please feel free to contact us!!

 

>>Contact Form

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 at TEL: 03-5321-3111 or by phone.

 

Related Articles

[Market Trends] Utilizing Neural Networks for Machine Translation

http://www.science.co.jp/mt/market/blog42.html

——————
★Machine Translation Seminar Video Now Available★

▼Please check the following video content page.
http://www.science.co.jp/mt/moviecontent.html
——————

 

Blog Writing Team

Tokuda Ai

・As a machine translation consultant, I provide consulting services for Japanese companies on machine translation implementation and process building.
・I place importance on the quality of the source text, which affects multilingual translation, and also provide consulting services for manual creation suitable for machine translation in the Japanese writing process.
・I also give presentations on the following topics related to machine translation.
- Presentation at the 23rd JTF (Japan Translation Federation) Translation Festival in 2013
"Approaches to Machine Translation in Multilingual Environments
- From the Perspectives of Evaluation and Process"
- Presentation at the 2014 AAMT (Asia-Pacific Association for Machine Translation) Machine Translation Fair
"Mastering Machine Translation - Improving Quality and Productivity"