XMATヘルプページ|LDX lab

Language Asset Management (Language Asset Registration)

作成者: Admin|Jul 24, 2026 12:01:31 AM

What is a language asset?

There are two types of language assets in this system: “TM (translation memory)” and “Glossary”. In both cases, as a general rule, the source and translation pair (TU = Translation Unit) is counted as one line. You can register (import) your language assets in the following three ways. For details, please see the section below for details. The registered language assets can be used in Quick PE and LAC (Custom MT Model Management) of this system.
(1) Import the language data (files) that you have locally.
(2) Extract language data from available texts on the Internet (or files you have locally) by leveraging AI, and import the aligned data.
(3) Import the stockdata provided in this system.
Service How to use TM
(Non Stockdata)
TM
(Stockdata)
Glossary Reference  file
Quick PE Select as a glossary     X  
Select as a translation memory X   X  
Use as a reference file X X X X
Quick MT  Use as a reference file X X X X
LAC
(Custom MT Model Management)
Create a custom trained model by using as corpus for the custom training X X X  
Create a custom glossary model by setting it to a general/custom MT model     X  

 

Information in Language Asset Registration screen

 You can register new language assets on this screen. Select the import category by referring to the table below depending on the format of the original file. Browse the file and set the languages, then click [Import] button to start the import process.

When importing language assets, you must manually specify the source and target languages. The language pairs of language assets supported by this system are English-Japanese/Japanese-English only. Specify the language according to the contents of the source file. 

 

How to register language assets (Import category)

The registration method (import category) of language assets is as follows. Specify the best method based on the type of original file and language asset. The initial value of the registered language asset name is the file name of the original file (excluding the extension).
*  indicates the recommendation degree.

Type
Import category Explanation Original file Supported
file format
Processing
time
Language asset
quality
TM TMX Upload
(*1)
Registers the translation memory (TMX) as it is as a TM. Since no conversion process is performed, high-quality language assets can be registered in a short processing time. Files in TMX format tmx *** ***
TM Conversion Creates a TM from a bilingual file (1 file) where the source and target text are paired. Bilingual files xliff、csv、tsv ** **
TM Alignment
(*2)
Creates a TM from monolingual files (2 files) for each of the source and target text. The alignment process takes a longer time, but you can create TMs from files of various formats. The source file and the target file should be in the same format. Monolingual files for source and target text docx、xlsx、 pptx、txt、html * **
TM Generation (AI)
(*2, 3)
Creates a TM from only the source monolingual file (1 file). Since translation is generated by the machine translation engine (*3), the processing time will be longer. Please choose this method when you have only the source file. You can import source files of 20,000 TU (lines) or less.
The resulting translation is not very accurate, so please edit the content in Language Asset Editor screen by clicking icon in [Edit] column of List screen.
Monolingual file for source text docx、xlsx、pptx、txt、html * *
Glossary Glossary Conversion
(*1)
Registers the contents of a glossary format file as it is as a glossary. High quality language assets can be registered in a shorter processing time. Glossary format files xlsx、csv、tsv、tbx *** ***
Bilingual Glossary Extraction Creates a glossary from a bilingual file (1 file) where the source and target text are paired. The processing time is longer because the process of extracting frequently occurring terms is performed.
The term extraction is not very accurate (for example, the source or target text of the term is missing), so please edit the content in Language Asset Editor screen by clicking icon in [Edit] column of List screen.
Bilingual files xliff、csv、tsv、tmx * *
Glossary Extraction Creates a glossary from only the source monolingual file (1 file). The processing time is longer because the process of extracting frequently occurring terms is performed. Please choose this method when you have only the source file.
The registered language asset does not include the translation, so please edit the content in Language Asset Editor screen by clicking icon in [Edit] column of List screen.
Monolingual file for source text docx、xlsx、pptx、txt、html、csv、tsv、srt、md、tmx、xliff * *
 

LeftoverTM and the criterion

When you register a TM as a language asset, with some import categories, you need to specify "Leftover criterion" for adopting TUs as a language asset from the original file. In this system, data that does not meet such standards is noise (mistranslation or correspondence deviation), and that is not suitable for TM as it is, then not adopted for TM. Therefore it will be treated as “Leftover”. “LeftoverTM” is a TM that collects only noise.

Depending on the data in the original file, LeftoverTM, which collects only data that does not meet the criterion, is registered separately in addition to the main TM that collects only the data that satisfies the criterion ("(LeftoverTM)" is automatically appended to the end of the registered language asset name). Please refer to the following to specify the criterion for adopting.

  • Strict (less noise): Since it is determined by a tight criterion, less data (number of TUs) is adopted, but noise is also reduced. Specify this if you want to create a high-quality TM from a file that contains a lot of noise.
  • Recommended: This is the default value. Please specify if it does not correspond to either “Strict” or “Relaxed”.
  • Relaxed (more noise): Since it is determined by a loose criterion, more data (number of TUs) is adopted, but the noise also increases accordingly. Please specify this if you want to adopt the data in the original file as much as possible into the TM.

Duplicate terms in the glossary

You can register duplicate terms (bilingual data) in the glossary, but it is recommended that a term pair (source-target text) be unique, whether referenced as a glossary or incorporated into a custom glossary model. However, there are cases where you want to change the translation of each term and the translation of a compound word, as shown below.
 English 
 Japanese 
asset 資産
management 管理
asset management アセットマネジメント

If you create a custom glossary model with a glossary which includes such bilingual data, when you machine translate the source text (in this case, English) using the model, the term with the longer source length will be prioritized and the translation will be adopted. In the above example, if “asset management” appears in the source text, “アセットマネジメント” will be prioritized as the translation, so this source text will not be output as “資産管理”.

 

About Kawamura NMT engine type

There are eight types of models (combinations of types and language pairs) for Kawamura NMT. The characteristics of each type are listed below. Please select the one that best suits your purpose.

  • General NT: Engine can be used for the translation in any field/purpose of the source documents.
  • Patents NT: Engine built specifically for patent translation, and supports specific terms and phrases.
  • Finance Services: Engine built specifically for financial services translation, and supports specific terms and phrases.
  • Finance NT (IR/Disclosure): Engine built specifically for the translation of IR/disclosure information related to financial service, and supports specific terms and phrases.
  • Legal documents NT: Engine built specifically for the translation of legal documents, and supports specific terms and phrases.
  • Science: Engine built specifically for translation in fields suitable for academic papers mainly in biological and medical fields, and supports specific terms and phrases.
  • Patent decision etc.: Engine built specifically for the translation of notices of reasons for refusal of patent applications and decision information, etc. among patent-related documents, and supports specific terms and phrases.
  • News: Engine built specifically for the translation of news, including new words, new topics and current topics, and supports specific terms and phrases.

Processing time of language asset registration

Please note that it takes a certain amount of time for the file imported in Language Asset Registration menu to be registered as a language asset. The processing time depends on the import category you choose, the size of the original file(s), and your network connection status. Once the status is “Import Complete” in Language Asset List, you can rename the language assets, edit, merge, or delete the source/target text in Language Asset Editor.