Everything you need to know about Translation Memory
Published on 06 May 12:43 by Erik Chan
What is Translation Memory (TM)
A translation memory (TM) is a dictionary of previously translated sentences. During the process of translating texts, translation memory is used to suggest identical or similar texts that were previously translated to help translate new texts more quickly and consistently. Translation memory is sometimes known as translation memory managers or translation memory systems in some software programs.
Translation memory is used by the majority of professional translators. In a survey of translators in 2006, 82.5 % of 874 translators confirmed using translation memory. TM is typically used in combination with a computer assisted translation (CAT) tool.
How does it work
Translation memory works at the sentence level. When a translator is working on translating a sentence using a CAT tool, the sentence is compared to sentences in the memory. If a similar sentence is found in the translation memory, the CAT tool will indicate this information to the translator and provide the option to:
i) use the TM translation,
ii) replace it with a new translation
or iii) modify the reference provided by the TM.
Sentences identical to those in the translation memory are called 100% matches. This means that the translation memory contains an example where the exact sentence has been translated before. It is even possible to have 101% and 102% matches, which means that not only the current sentence is 100% the same, but also one or both of those sentences before and/or after it are the same. Sentence matches that are similar but not 100% the same are called fuzzy matches. Matches are usually ranked from 0% to 99% based on the similarity. A 99% match means that the segments differ by one character. Matches below 70% are generally not useful and are ignored by the translator.
Translation memory is most beneficial when the texts to be translated are highly repetitive, such as fund documents or legal contracts. They are also useful when translating incremental changes in a previously translated document. Translation memories are much less useful for creative or literary texts because repetition is less frequent (it is unlikely the translator has translated the same sentence previously). However, some translators find translation memory valuable even for non-repetitive texts, because the memory can help provide reference to determine appropriate usage of particular terms and for quicker review.
The technology behind Translation Memory
Translation memory is generally stored in a file with each sentence pair (the source text and translated text) listed one at a time. While there are a number of different file types to store translation memory (TMX, TBX, UTX, SRX, GMX, OLIF, XLIFF, TransWS, xml:tm, PO), they are all the similar and easily interchangeable since it is simply really just a list of dictionary items.
When a translation memory is being put into use, each sentence pair in the memory is compared with the sentence to be translated to determine their similarity. In computer science, this is called approximate string matching (often referred to as fuzzy string searching). Approximate string matching is typically divided into two components:
i) finding approximate word matches inside a given sentence
and ii) finding sentences in the translation memory matching the sentence approximately.
The University of Auckland has a great presentation on the various string matching algorithms here.
How useful is translation memory?
The idea behind leveraging a translation memory is simple. It saves translators time translating the same sentences repeatedly, and with this comes a list of benefits:
Consistent Translations: When the same sentences need to be translated again and again, translation memory suggests previous translations to keep consistency throughout all the work.
Centralized Database: When one or more translators are working on the same texts, they will more likely use the same translation based on the references and suggestions by the translation memory.
Cost Savings: Time savings for translators translates to cost savings. Moreover, translation memory is like an accumulated investment where the more translated sentence pairs it stores, the more quickly the translator can translate new (similar) texts.
Improved Quality: Editors, managers, and other translators can refine and optimize translation memory entries to adhere to a consistent tone and brand voice.
There are claims that using translation memory can lead to an increase in productivity between a wide range of 15% to 90%. The range is largely dependent on the TM content as suggested by Masaru Yamada of Kansai University's Faculty of Foreign Language Studies (FFLS) in his research on the effect of translation memory databases on productivity. In a separate study Memsource demonstrates that the average increase in productivity via translation memory usage is around 36%
Other things you need to know
It is important to note that translation memory is more beneficial for keeping the consistency of whole sentences and less so for consistent terminology usage. Glossary management is generally a much better solution to resolve terminology inconsistency and will be discussed in my next post. This is due to the fact fuzzy string matching algorithms take into account entire sentences and are not great algorithms encompassing terminology differences when comparing two sentences.
Some translation service providers will offer a discount on sentences already in a translation memory when asked (yes, it is a dirty industry secret). Discounts are often provided based on the similarity of the sentences translated to a match in the translation memory. 100% matches are sometimes called ICE (in-context exact) matches, and anything below 100% is called a fuzzy match.
Key Takeaway to Share:
- What is Neural Machine Translation & How does it work?
- What is Neural Machine Translation & How does it work?
- What is a Computer-Assisted Translation (CAT) tool
- Top 3 challenges translating equity research reports
- This is how long it takes to translate financial documents
- Keys Factors in Quality financial translation
- Best sites to hire professional translators
- How much to pay financial or legal (professional) translators?
- Add AI language capabilities to increase cross-border business transactions
- 3 reasons investments banks should adopt AI translation tools now
- Reshaping equity research business with AI
- With the rise of AI translation, why hire human translators
- Is AI Ready to Translate Financial News
- AI translation for Company Financial Reports
- Applying AI translation to Equity Research Reports
- Does my company need a translation management system
- Should my company implement AI translation tools
- AI Machine Translation and Terminology Consistency
- Security & Compliance of AI Machine Translation: What you need to know
- Everything you need to know about Translation Memory
TranslateFX develops AI translation technology specifically for financial and legal institutions. The company develops AI models and workflow tools for clients of all sizes. We believe humans always play and important part of the process and our tools reduce the time and costs of translation by 60% or more.
- China Securities Regulatory Commission
- China Banking and Insurance Regulatory Commission
- China Banking Regulatory Commission
- People's Bank of China
- U.S. Securities and Exchange Commission
- U.S. Financial Industry Regulatory Authority
- U.S. Financial Accounting Standards Board
- Hong Kong Securities and Futures Commission