Completely confused about MT categories and impact on the model

Question

I am trying to train an engine in the Microsoft Hub while taking advantage of the significantly better Neural vs SMT technology. While testing on https://translator.microsoft.com/neural/ , it's clear to me that Neural is what I want.

However, the documentation of the hub at //hub.microsofttranslator.com/Help/Download/Microsoft%20Translator%20Hub%20User%20Guide.pdf is confusing:

On page 7, paragraph 2.3.1 it says

The category identifies the domain – the general area of terminology you want to use – for your project. Please choose a category that is most appropriate and relevant to your type of documents.

In some cases, your choice of the category directly influences the behavior of the Hub:

• If you choose “Technology” and choose to use Microsoft models, the Hub will use a different set of underlying models than for all other categories.
• If you choose “Speech”, the Hub will use models that are optimized for processing the output of speech recognition systems.
• Any other category selection uses the “general” Microsoft models, and is used as an identifier in your category ID.

But then on page 51, it completely contradicts that statement and says:

Q: I selected ‘Technology’ category when creating a project. When I train the system for this project and check the option “Use Microsoft Models”, will the training use Microsoft’s Technology model?

A: No. At this moment the choice of “Use Microsoft models” always invokes the general training data, same as Bing Translator. The category selection serves only to identify your purpose, but has, as of now, no influence on the behavior during training or translation. That will change in an upcoming release.

Can someone sort this out? What mix of Category and "use MS model" option do I have to pick to get the "good" Neural MT output I'm looking for?

Thanks!

@Chris_Wendt wrote this recently in another post:

Using the parameter &category=generalnn in the API call will produce the same results as you see in the Bing UI. Be careful with tagged content. Tags are handled better with the default category value, blank or "general".

I'm using the SDL Trados 2017 plugin for Microsoft translator. Do I have a way to pass this parameter?

[

Answer 1

The neural systems are now customizable using the new preview of Custom Translator. You need to be using the Microsoft Translator Text API v3 and Custom Translator supports language pairs that are NMT languages.

Answer 2

You can pass "generalnn" or your Hub customized system's ID in the Category ID field in Trados, or any other TM tool that lets you set the category value. The neural systems are not customizable at this point, but you can use it the same way you would use a customized system.