XTM Cloud v 13.1

General settings

Administrators must connect a machine translation engine so that linguists can benefit from machine translation matches. During the setup, Administrators can decide whether to insert machine translation matches directly into the target segments, ready for post-MT editing.

Connecting Machine Translation engines

For detailed connect a custom Machine Translation engine.

  1. Go to Configuration > Settings > Translation > Machine translation > MT engines.

  2. Select the required machine translation engine.

  3. In the General options select:

    1. Do not search for an MT match when one of the selected matches exists

    2. Populate target with match

    3. Set MT matches to done

  4. Optionally, for selected engines, you can choose when to Send text for matching: Only for these engines:

    • Google Translate

    • Microsoft Translator

    • Amazon Translate

    • Amplexor MT

    • SYSTRAN MT

    • DeepL Translator

    • BeGlobal MT

    • Iconic MT

    • SAP Translation Hub

  5. Once you have selected your MT engine, decide how you want to have your text analyzed:

    1. In XTM Workbench – the text is sent to an MT only when a user enters a segment in XTM Workbench in the Edit mode. The translation is provided on a segment-to-segment basis.

    2. After analysis – the whole text is sent to MT matching during project analysis. When the user enters XTM Workbench, all segments are filled out with MT matches

    3. After analysis and in XTM Workbench – the whole text is sent to MT matching during project analysis. The text can be resent to the MT on a segment-to-segment basis as in “In XTM Workbench”.

  6. Select Save. 

All segments without TM matches are sent for MT matching. You can also send segments that have varying levels of TM matches.

Amazon Translate

The machine translation engine details are available here.

Amazon Translate is a neural machine translation (NMT) service that delivers fast, high-quality, and affordable language translation. Custom Terminology is a feature used to customize Amazon Translate output to use company- and domain-specific vocabulary. By uploading and invoking Custom Terminology with translation requests, you ensure that your unique content, such as brand names, character names, and model names, is translated exactly the way you need it.

Amplexor (Acolad)

The machine translation engine details are available here.

CrossLang

The machine translation engine details are available here.

Custom Machine Translation

XTM provides the option to enable and configure the connection to a custom MT engine.

How to add custom MT

Click Add custom MT to open a configuration window.

  1. Fill out the required fields:

  1. Enter a unique name for your Custom MT.

  2. Provide the username and password used to log in to MT.

  3. Enter API Base URL.

  4. Provide your endpoints for authentication and translation.

  5. Provide the Check status endpoint and time intervals at which the status has to be checked.

  6. Provide your endpoint for downloading translation.

  7. Choose custom fields (which are configurable for each MT engine – see more on project custom fields here).

  8. Optionally, provide a prefix of your custom field (to distinguish custom fields metadata sent to MT from other metadata).

  1. Click Add.

The decision whether to use Custom MT is configurable:

When creating a project, in the Project Editor, and in the Project template editor.

Connecting custom Machine Translation engine
Connecting XTM account to your custom Machine Translation engine

  1. Go to Configuration> Settings > Translation > Machine Translation.

  2. In the MT engines, select the Add custom MT button.

  3. A new window comes up. Fill in all the required details marked with a red asterisk asterisk-solid.png.

    - Custom MT name

    - Username

    - Password

    - Base URL (API Base link)

    - Authentication endpoint

    - Check status endpoint

    - At time intervals (at which the status has to be checked, in seconds)

    - Download translation endpoint

    Optionally, you can also choose:

    - Custom fields (which are configurable for each MT engine).

    - Custom fields prefix (to distinguish custom fields metadata sent to MT from other metadata).

  4. Select Add.

The new MT engine is displayed in the Configuration > Settings > Translation > Machine Translation section. You can add it when you create a project or in the Project Editor.

DeepL

The machine translation engine details are available here.

XTM users can implement following optional DeepL parameters:

  • formality — description is available here under formality

  • glossaries — description is available here under glossary_id and detailed instructions on creating and managing DeepL glossaries is available here

To enable DeepL glossaries in XTM:

  1. Go to Configuration > Settings > Translation > Machine Translation > DeepL Translator section

  2. Fill in the Authentication key field

  3. Select Synchronize

As a result:

A green label Synchronization successful and a new parameter Glossaries appear.

Note

When a synchronization error occurs, the following error message appears: Cannot synchronize data with DeepL Translator. Check all connection details. Make sure that you use a correct Authentication key and that your DeepL glossaries and subscription are valid and configured properly.

To see the list of the successfully synchronized DeepL glossaries:

  1. Go to Configuration > Settings > Translation > Machine Translation > DeepL Translator section

  2. Select View next to the Glossaries parameter.

As a result:

The DeepL glossaries popup will appear showing the glossary information:

  • name

  • source language

  • target language

Google Machine translation

By default, Google machine translation is switched off. This is to ensure the security and confidentiality of your data. While your data is kept on the XTM server it is secure, however when it goes out to be processed by Google Machine translation, XTM International cannot guarantee its confidentiality. We recommend that you only switch this option on if you are not working on confidential information.

Texts can be sent for Google machine translation after the project analysis when working in the XTM Workbench or both.

The Google Translate API v2 is a paid service that is not covered by your XTM Cloud subscription, therefore if you decide to use this option, you need to create an API key and have a paid account with Google.

Check Google Developers Console to generate API key.

Check Google Cloud Translation pricing page for info on pricing.

Iconic Translation Machines

This is a neural MT (NMT) engine provider. More about Iconic MT can be found here.

Intento

For configuration, contact us at support@xtm-intl.com ()

KantanMT

Default setting: This is the default setting whether KantanMT should be used for new customers and for new projects.

API URL: This is the endpoint of the KantanMT server. The default value is: http://www.kantanmt.com/api.

Token: This is the password to KantanMT's API and should be provided by KantanMT. This value is passed to the KantanMT server during translation, and it is not sent to translators.

Lingo24

The machine translation engine details are available here.

Lionbridge GeoFluent

The machine translation engine details are available here.

Microsoft Translator

You can find more details here: https://www.microsoft.com/en-us/translator/default.aspx.

There are 3 ways to access Microsoft Translator:

  1. XTM shared account: free of charge. The account is not private, so no Client IDs or Client Secrets are required.

  2. Azure Marketplace: If you do not use the XTM shared account, you need to get your token to access Microsoft Translator API. Learn more here: https://docs.microsoft.com/en-us/azure/cognitive-services/translator/translator-text-how-to-signup

  3. Microsoft Custom Translator:

    Custom Translator creates a private memory for a specific key. You can train this memory with domain-specific TMs. To find out more about this solution and sign up, go to: https://portal.customtranslator.azure.ai/ .

Omniscien Language Studio

XTM has been integrated directly with the Language Studio translation and training platform using the Language Studio API.

Before setting up XTM to work with Language Studio, you create an account with Omniscien. Then using the information provided from Omniscien complete the fields in this configuration section. You can then set within each customer record whether you wish to use Language Studio or when creating a new project, you have the opportunity to switch on Language Studio and change the default settings for the Language Studio project and Domain.

The integration works in the following way: When you create a new project, XTM analyses the source file and matches the segments with your TM. XTM then sends the XLIFF to Language studio for matching of any unmatched segments. This may take some time so the project creation will take longer than is normally the case. Once complete, Language Studio sends the matched XLIFF back to XTM and the file is ready for translation in XTM Workbench.

Oneliner

The machine translation engine details are available here.

SAP Translation Hub

This integration is based on the SAP Document Translation service. The documentation can be accessed here

SDL BeGlobal

A cloud-based machine translation solution, integrated with SDL BeGlobal API for customization.

More information about API authentication

SmartMATE

Provided by Capita TI. You can find more details here.

SYSTRAN Pure Neural™

The machine translation engine details are available here.

Thanks to the Neural Fuzzy Augmented powered by SYSTRAN, it is possible to improve machine translation matches using the highest fuzzy matches coming from the translation memory up to the 84% threshold.

For the Neural Fuzzy Augmented, SYSTRAN offers two types of accounts:

  • Free translation of 1.000.000 characters per calendar month

  • Pro unlimited translation

Enabling Neural Fuzzy Augmented

  1. Go to Configuration > Settings > Translation > Machine translation > General options. Deselect the 75-84% fuzzy checkbox.

  2. In the SYSTRAN MT tab, select the Enable Neural Fuzzy Augmented matches checkbox.

  3. Select when to send text for matching:

    1. In XTM Workbench

    2. After analysis

    3. After analysis and in XTM Workbench

A fuzzy machine translation match based on an existing fuzzy translation can be accessed in the Matches panel in XTM Workbench if:

  • the maximum level of fuzzy match is 84%

  • there is no higher match than fuzzy in the segment (ICE, leverage, repetition)

  • there is at least one fuzzy match in the 75-84% range

Tauyou

tauyou <language technology> provides domain-specific machine translation solutions for LSPs. With the private Translation Memories (TMs) for each client, and additional public domain data, tauyou creates a dedicated system for every customer, including translation glossaries and forbidden word lists. tauyou has experience in 30+ languages (and their combinations).

For more information, and to set up an account please send an email to info@tauyou.com (info@tauyou.com).

In this section:

is an approach to machine translation that uses an artificial neural network to predict the likelihood of a sequence of words, typically modeling entire sentences in a single integrated model.

source: Wiki

is an approach to machine translation that uses an artificial neural network to predict the likelihood of a sequence of words, typically modeling entire sentences in a single integrated model.

source: Wiki