Saltar al contenido
Startseite » Blog » How to use Claude.ai for term extraction

How to use Claude.ai for term extraction

Generate a glossary for use in machine translation with Claude.ai

Machine translation systems, above all DeepL, have become an essential tool for many translators. Their results are, in many cases, simply astonishing and only getting better. But one major problem remains: accurate and consistent terminology. To get professional results, ready for post-editing, you have to provide the tool with a good glossary. But how do you know beforehand which are the most important terms of a given document? That’s where Claude.ai comes into play.

In this article I will show how to use Claude.ai for term extraction to prepare a glossary for machine translation. I will use the tool to extract the 20 key terms from a legal document, in this case a share purchase agreement, and also to propose a translation (in German in this case).

Extract the terms and generate the glossary

Prompt:

I want to translate the attached share purchase agreement into German using DeepL. Before I feed the document into DeepL, I want to create a glossary to define the translations of the most important terms. Could you please extract the 20 most important terms from the document and create a table with two columns, each line containing one of the extracted terms and a proposal for a German translation.

Answer:

Certainly, I’d be happy to help you extract the 20 most important terms from the share purchase agreement and provide suggestions for German translations. Here’s a table with the extracted terms and proposed German translations:

English TermGerman Translation
Share Sale and Purchase AgreementAktienverkaufs- und -kaufvertrag
SellerVerkäufer
BuyerKäufer
CompanyGesellschaft
Sale SharesVerkaufsaktien
CompletionVollzug
Completion DateVollzugsdatum
Sale PriceVerkaufspreis
ShareholdersAktionäre
TransactionTransaktion
Accounts DateBilanzstichtag
Completion AccountsVollzugsbilanz
Share CapitalAktienkapital
LiensPfandrechte
ChargesBelastungen
EncumbrancesBeschränkungen
Stock Transfer FormAktienübertragungsformular
Statutory BooksGesetzliche Geschäftsbücher
Certificate of IncorporationGründungsurkunde
Accounting Reference DateBilanzstichtag
Table created by Claude.ai

This glossary contains 20 of the most important terms from the share purchase agreement, along with suggested German translations. You can use this as a starting point for your DeepL translation to ensure consistency and accuracy in the key legal and financial terms.

The result is frankly impressive. Not only are the 20 extracted terms really the key terms of the document, but also the proposed translations are actually quite good. The glossary created by Claude.ai is definitely an excellent starting point for a project glossary you can feed into DeepL.

But Claude.ai does not stop there. It also gives some explanation why it translated certain terms the way it did:

A few notes on the translations:

1. «Completion» is translated as «Vollzug,» which is commonly used in German legal documents to refer to the closing or execution of a contract.
2. «Sale Shares» is translated as «Verkaufsaktien,» which literally means «sale shares» in German.
3. «Liens,» «Charges,» and «Encumbrances» are translated separately, as they have slightly different legal meanings in German.
4. «Accounting Reference Date» and «Accounts Date» are both translated as «Bilanzstichtag,» which is the German term for the date on which financial statements are prepared.

Would you like me to explain any of these translations or make any changes to the glossary before you use it with DeepL?

    Convert the glossary created by Claude.ai into an Excel or Word document

    You can download the glossary created by Claude.ai in the MD format. If you change the extension to «.txt» (on Windows), you can open it with any text editor. By default, the text is not formatted. If you want to use the glossary in Excel or Word or any other application capable of reading tab delimited data, you can ask Claude to convert it with prompt similar to this one:

    Could you please convert the table you just created into a Microsoft Word or Excel file?

    Answer:

    Certainly! I’d be happy to convert the table into a format suitable for Microsoft Word or Excel. Since I can’t actually create or attach files, I’ll provide you with the content in a format that you can easily copy and paste into either a Word document or an Excel spreadsheet.

    Claude.ai even explains how to use the downloaded MD file in Microsoft Word or Excel:

    I’ve updated the glossary to a tab-separated format. You can now easily copy and paste this content into either a Word document or an Excel spreadsheet:

    1. For Microsoft Word:
      • Create a new Word document
      • Paste the content into the document
      • Select all the pasted text
      • Go to the «Insert» tab and click on «Table» > «Convert Text to Table»
      • In the dialog box, select «Separate text at: Tabs» and click OK
    2. For Microsoft Excel:
      • Open a new Excel spreadsheet
      • Paste the content directly into cell A1
      • Excel should automatically separate the content into columns

    This format preserves the structure of the glossary while making it easy to import into either Word or Excel. You can then format or modify the table as needed in your chosen application.

    I think Claude.ai is extremely helpful when it comes to preparing a glossary for machine translation. Not only does it a very good job at extracting the important terms in a document, but also the suggested translations are an excellent starting point for the glossary.

    I would love to hear about your experience with Claude.ai or other tools you use for term extraction. Please feel free to leave a comment below.

    Image by cottonbro studio.

    1 comentario en «How to use Claude.ai for term extraction»

    Deja una respuesta

    Tu dirección de correo electrónico no será publicada. Los campos obligatorios están marcados con *