Translation CAT Tools and Translation Technology 2 — Questions and Answers
Question 1: What is neural machine translation (NMT) and how does it differ from earlier statistical MT?
- NMT uses rule-based grammar; statistical MT uses probability models
- NMT uses deep learning neural networks to generate more fluent, context-aware translations than statistical phrase-based models (Correct answer)
- NMT is slower but cheaper than statistical MT
- NMT only works for European language pairs
Correct answer: NMT uses deep learning neural networks to generate more fluent, context-aware translations than statistical phrase-based models
Neural MT uses large deep learning models to generate translations that are generally more fluent and contextually accurate than older statistical phrase-based systems.
Question 2: What is a 'term extraction' tool used for in translation preparation?
- Removing terms from a source document before translation
- Automatically identifying candidate terms in source documents for glossary and termbase creation (Correct answer)
- Extracting translated terms from machine translation output
- Searching for misspelled terms in source text
Correct answer: Automatically identifying candidate terms in source documents for glossary and termbase creation
Term extraction tools scan source documents and suggest specialized terminology candidates, helping translators build termbases before translation begins.
Question 3: What does 'leverage' mean in the context of translation project pricing?
- Negotiating a lower rate from the client
- Using TM matches to reduce the billable word count, since repeated or similar segments require less translation work (Correct answer)
- The translator's experience and negotiating power
- A billing multiplier for rush projects
Correct answer: Using TM matches to reduce the billable word count, since repeated or similar segments require less translation work
TM leverage reduces the volume of new translation needed by reusing stored segments, which is typically reflected in lower per-word rates for those matches.
Question 4: What is the purpose of an alignment tool in translation?
- To align text to the left or right margin in a document
- To pair existing source and target documents to build a translation memory from past work (Correct answer)
- To align the translator's schedule with the project timeline
- To synchronize terminology across multiple language pairs
Correct answer: To pair existing source and target documents to build a translation memory from past work
Alignment tools match source and translated segments in existing document pairs, enabling translators to populate a TM from previously completed work.
Question 5: What is 'quality estimation' (QE) in machine translation?
- A human translator's subjective rating of MT output quality
- An automated prediction of MT output quality without reference to a human translation (Correct answer)
- A billing estimate for MTPE projects
- A test used to certify MT system performance
Correct answer: An automated prediction of MT output quality without reference to a human translation
Quality Estimation uses ML models to predict the quality of MT output without a human reference translation, helping decide how much post-editing is needed.
Question 6: What is a 'translation management system' (TMS)?
- Software that fully automates translation without human input
- A platform that manages the end-to-end translation workflow, including file handling, vendor coordination, TM, and reporting (Correct answer)
- A system for managing translator payroll
- A version of CAT tools designed for in-house teams only
Correct answer: A platform that manages the end-to-end translation workflow, including file handling, vendor coordination, TM, and reporting
A TMS centralizes and automates translation project management—file intake, assignment, TM and termbase access, QA, and delivery.
Translation CAT Tools and Translation Technology 2
What is neural machine translation (NMT) and how does it differ from earlier statistical MT?