Improve translation quality and speed with this quick how-to

Machine translation – the automatic translation of texts using computer software – has been embraced by both companies and individuals in the last few years. The main advantage is that it takes next to no time, regardless of the volume of the text translated. And there are more benefits:

• Lower cost. Machine translation can be up to 100 times cheaper than translations done by humans.

• Secure data. Texts are transferred between the client and the machine translation system in an encrypted form. In most cases, they only remain in the memory of the machine translation engine for a short time and are not used for any other purpose.

• Consistency. All terms and repeated phrases are translated the same way to ensure the text is uniform.

In the past, machine translation has been dismissed as a limited tool, used mainly for texts with a short life cycle or where quality isn’t critical, such as very basic emails or product descriptions and reviews. But machine translation technology has improved dramatically in recent times, making it a powerful tool in almost all translation scenarios – as a stand-alone translation tool or as one of many components within a translation workflow that also includes professional translators.

In this article, we’ll show you how to make the most of machine translation tools, and we’ll focus on speed and quality in particular.

How machine translation works

There are different types of machine translation (MT), such as rule-based, statistical, and example-based. We’ll focus on neural machine translation and artificial intelligence –  the two types that have produced the best results, even competing with human translations.

Neural Machine Translation

Neural machine translation (NMT) is based on deep neural networks. A network is made of two components: an encoder that reads the input sentence and generates a representation suitable for translation, and a decoder that generates the actual translation. The evaluation of the fluency of the resulting sentences is what allows NMT to be more grammatically accurate than previous generations of MT.

Artificial Intelligence

Machine translation engines continuously collect and sort data to improve output both in general and specific domains. This data is used to train the engine so that it can choose the best translation option based on the context.

More texts translated fast

So, let’s start with speed. The beauty of machine translation is that it translates text almost immediately. You can also translate a text into several languages at once.

What you get

• Translating as much as you like at the same time means you can quickly adapt your materials for use in different countries at the same time – this is great for elearning courses, for example.

• If you want to edit your texts after translation, you can quickly have all translations ready for your team to start working on them and speed things up.

• It’s cost-effective: with our Smartwords technology, for example, you can afford to translate content that you just wouldn’t be able to translate any other way, given the constraints of traditional forms of translation (time, cost of labor, etc.).

How to do it in Smartcat

Create a new translation,

1. Drag and drop your files,

2. Choose the source and target languages,

3. Hit the ‘Translate files’ button and proceed to edit or download the translations.

Editing gets easier every time

No translation is perfect so a little editing can go a long way, especially when working on automatically translated texts. How much editing you’ll need depends on the type and the purpose of the text. That's why most CAT (Computer Assisted Translation) tools, including Smartcat’s, are equipped with an integrated editor that helps speed up the process.

The best thing about these editors is that corrections can be saved as translation memories and applied to any similar segments that you’re translating or that you may translate in the future.

What you get

• You can work on the text alone or in parallel with colleagues without sacrificing consistency.*

• No need to go searching through the text every time you want to apply the same or similar edits.

• No need to worry about whether you have already translated a specific sentence or passage.

How to do it in Smartcat

Open your translation in the Editor.

1. Check, edit, and approve each segment. Sentences with potential errors and inconsistencies are highlighted with a warning sign.

2. Approve all changes.

3. Download the resulting file.

*The version history, questions, and comments are available at the bottom of the Editor screen.

However, there are more ways to improve your text. To see all options click the corresponding button in your translation pack.

You’ll be shown all the ways you can process your text: edit it yourself, invite your colleagues and work together, or hire a linguist from the Smartcat Marketplace* to help you.

*Please note that you need to create a translation project to work with freelancers and agencies. Click the ‘Improve’ button to switch to the ‘Manager view’ described in our ‘Getting Started Guide’.

Save on human resources

Human translators and editors can be pricey and may not always be readily available. That's why reducing the time and effort they have to put in – without sacrificing translation quality – is a huge benefit.

A good CAT tool will help you produce translations that require minimum editing, but that doesn’t mean you can’t leverage machine translation for use with professional translators and editors. It can significantly reduce manual work and increase productivity.

What you get

• The whole team has access to machine translation whenever they need it,

• Everyone knows what stage the project is at at any time,

• You can easily estimate if you need more people working on any given task.

How to do that in Smartcat

More hacks to get quicker, better, and cheaper translations

If you’re ready to try a CAT tool to streamline your translation process, here are more tips for you.

Choose the language pair wisely. If you need to translate a text into several languages, it’s good to keep in mind how they relate to one another, to help you minimize the editing required. For example, if you need to translate a text from Russian into English and German, it might make more sense to translate it into English first, edit the translation, and then translate it from English to German.

Prepare the text. If you’re creating a text to be translated from scratch, try to keep the wording and structure straightforward to make the most of the CAT tool: use short, simple sentences, avoid slang, jargon, and ambiguous language, reduce the number of adverbs, etc.

Stick to the right CAT tool. The more you translate, the more the CAT tool learns from your texts, and the better the translations will be. So, ideally, once you find the tool that best works for you and your team, keep using it to increase its efficiency!

We hope this article opened up a whole new outlook on machine translation and what it can do for you. By carefully analyzing your translation needs and coming up with a good machine-plus-human combination, your new workflow could save you considerable time and money.