Phrase and beyond
Translation Technology: A Look at Its Past, Present, and Future
Translation technology has been around for more than 50 years now, but as our world has become increasingly interconnected, it’s only grown more essential.
A report by Nimdzi Insights found that the current language technology landscape includes more than 700 solutions. Given this diversity, how do you find a translation tool that best suits your business needs?
The first step is to gain a solid understanding of what translation technology is and how you can leverage it. Keep reading to find out.
What is translation technology?
Translation technology is the use of software tools that support the process of converting written text from one language to another. As is the case of most work aided by technology, translation technology tools can increase productivity, accuracy, and overall effectiveness.
The term “translation technology” encompasses everything from translation memories and terminology management tools to machine translation (MT), virtual interpreting technology, and even voice-to-text technologies.
Before the emergence of translation technology, translation was done manually, with translators consulting paper dictionaries and using their best judgment.
The negative impact on business was substantial as a result of:
- Delayed time-to-market
- A general loss of consistency across content
- The high costs of inefficient operations
- Lower output quality resulting from having to manually check for errors
The birth of translation technology changed it all. Tools such as translation memories—which store previous translations to inform and expedite future projects—and translation management systems help automate repetitive processes and increase control.
How does translation technology help businesses?
Today, translation technology tools allow businesses to:
- Ensure quality by checking for spelling and grammar errors
- Increase consistency across content by storing previous translations that can be reused or consulted in future projects
- Improve operational efficiency by automating translation management tasks
- Accelerate time-to-market by enabling the production of more content in less time
Before we go into more depth on the various types of translation technology, let’s go back in time and examine how translation technology evolved.
The history of translation technology
Some of the techniques used in modern translation technology can be traced back to the 9th century when an Arabic cryptographer named Al-Kindi developed the method of frequency analysis that is still used today.
However, it wasn’t until the mid-20th century, when computers became available and affordable, that translation technology truly began to take shape.
Here’s a rough overview of the evolution of translation technology:
- 1950s: Georgetown University and IBM introduced the world’s first machine translation (MT) system. The approach was rule-based and lexicographical, which means that it relied on pre-programmed rules and dictionaries. Although this early form of MT proved unreliable and slow, it was still revolutionary—a stepping stone on the path toward more advanced technology.
- 1970s: The United States Department of Defense and Defense Advanced Research Projects Agency (DARPA) started developing speech recognition technologies that paved the way for voice-to-text technologies.
- 1980s: The arrival of electronic dictionaries and terminological databases during this decade was another major turning point. These tools helped to make translation more accessible by providing translators with instant access to information (terminology with its translation) that could be used during the project.
- Mid-1980s: The precursors of modern translation management systems (TMS) entered the scene from the hand of Coventry Lanchester Polytechnic University and its ALP System.
- Late 1980s – early 1990s: IBM researchers introduced statistical machine translation (SMT). These systems were word-based and trained to translate one language into another by comparing large amounts of parallel texts in both languages (bilingual corpora). For example, they would analyze how often the German phrase “das auto” was translated as “the car” vs “the vehicle” vs “the automobile”, and choose the most frequent translation for the text at hand.
- Early 1990s: Most commercial computer-assisted (or aided) translation (CAT) tools appeared during this decade—a milestone that transformed translation technology forever. It enabled a whole new generation of translators to work more efficiently and effectively.
- Late 1990s: A new version of IBM’s statistical translation engine, this time phrase-based instead of word-based, was released. It became the commercial standard for years to come until Google entered the fray in 2006 with their neural machine translation (NMT) technology.
- Early 2000s: The first cloud-based TMS solutions appeared in the market, enabling translation teams to work more flexibly and collaborate with other company members regardless of location.
- 2006: Google launched Google Translate—still statistical—which took the world by storm. The system first translated the input text into English before translating it into the target language. The system used predictive algorithms, which would guess which words should come next, based on the words and phrases it had “learned” before. These guesses often resulted in poor grammatical accuracy.
- 2016: Google Translate introduced neural machine translation (NMT), which outperformed phrase-based CAT tools and became the new commercial standard.
Why is translation technology important in today’s globalized world?
In this fast-paced, ever-connected world, consumers demand seamless experiences that are as user-friendly and accessible as possible. They also want high-quality products and services that are culturally sensitive and adapted to their needs.
The process of adapting a brand’s message, look, and feel to a given culture or location is called localization. While localization may involve different aspects, such as product packaging, web design, or business strategy, it’s the culturally aware translation of a brand’s message into a specific language that is at the core of successful localization.
In the context of the information economy, this can mean having to deliver dozens of content types, in multiple languages, to a diverse range of audiences. Translation technology enables businesses not only to meet these challenges but also to maximize translation spend by increasing speed and quality while reducing cost.
Today’s decision makers who have to strike a balance between the expectations and demands that come with globalization while keeping a close eye on costs and productivity can rely on translation technology to help them deliver on all fronts.
Translation technology allows companies to go global quickly, easily, and affordably by enabling them to streamline the entire localization process. It provides them with scalable, efficient solutions that can help them reach new markets faster and more effectively than ever before.
The main types of translation technology
Modern translation technology is constantly evolving, with new innovations coming every year. What follows is a (non-exhaustive) list of the most popular translation technologies.
Computer-assisted translation (CAT) tools
CAT tools are software applications created to support translators in their daily work. CAT tools use databases of previous translations (usually from a specific source language and target language) as well as frequency information, segmentation data, and a wealth of other resources to aid the translation process.
Visually, a CAT tool interface usually displays two columns or panels: One shows the source text, and the other displays the target text. Typically, the tool will segment the source text into chunks—sentences or paragraphs—to make it easier for the translator to keep track of the flow of information.
CAT tools usually have a large number of functions, such as:
- Translation memories, which allows translators to reuse previous translations
- Term bases or glossaries, which can be used to search for brand-specific or project-specific terminology
- The option to click through to secondary resources for additional context, such as images or comments left by other translators, the project manager, or other company stakeholders
- Advanced search and navigation tools
- Reports detailing how much of the translation has been completed
- Auto-completion of segments if they are an exact match to previously translated content
- Quality assurance tools to search for errors in the translation, such as untranslated segments, missing numbers, or instances where the same word has two different translations
- The possibility of generating a final translated document that automatically mirrors the original document’s format
- Integration with other translation technologies, such as neural machine translation engines
CAT tools have revolutionized translation technology, making it easier for companies to optimize the translation process and handle large amounts of content in a more efficient way, saving both time and money. For translators, the main benefit is that they can use the time saved on repetitive tasks to focus more on the translation itself.
CAT tools are also highly scalable and customizable, making them invaluable for any business seeking to streamline its translation process based on the specific needs of its content and target audience.
Machine translation (MT)
Machine translation is an automated system that allows for the production of translated content without human intervention. As described earlier in this guide, machine translation is not a new phenomenon. However, the advances in technology that have taken place in recent years have led to a rise in the use of MT. Machine translation differs from CAT tools in that it doesn’t rely on human input but produces translations entirely on its own.
Types of machine translation
Since their inception, machine translation engines fall in one of the three below categories:
- Rule-based machine translation (RBMT): The engine translates words and sentences based on a set of rules that can be adjusted by the user.
- Statistical machine translation (SMT): The system translates based on a corpus (a large body of parallel texts) that the engine has been trained on.
- Neural machine translation (NMT): The engine is AI-powered and capable of mimicking the way a human brain works when processing language. In a nutshell, thanks to machine learning technology, these engines are capable of understanding entire sentences and even paragraphs at once instead of translating word by word.
Until 2016, MT systems were based on a statistical approach. Today, most machine translation engines are neural.
Raw vs post-edited machine translation
The quality of machine translation output varies depending on factors, among which we count:
- The language pair at hand: Languages of dissimilar syntax such as English and Chinese may be harder to translate accurately than languages with similar grammars, such as Spanish and French.
- The subject matter of the original text: A general-language document tends to translate more accurately than, for example, a legal text.
- Whether the source text undergoes pre-editing to eliminate certain elements that may cause the translation engine to struggle.
- The scope of the project and the amount of data available for training the engine.
- The use of editing resources or human translators to refine the translation output.
- The provider: Some engines perform better than others as a result of having access to more data.
Considering the above, companies can choose whether to use raw machine translation (i.e., completely unedited) or to hire human translators to refine the output, depending on their goals and needs. The latter is called “machine translation post-editing (MTPE),” and it can be either light or full.
Remember: For raw machine translation, it’s vital to use a state-of-the-art MT engine that is highly trained and powered by AI technology.
Thanks to modern technology, it’s also now possible to identify the quality of machine translation output to focus post-editing resources where they are most necessary. This is called machine translation quality estimation (MTQE).
Different types of texts will be more suitable for a lower level of editing. For example, a website’s FAQ page may require light editing; however, a press release that will directly impact the brand’s reputation should go through a full post-editing process.
By the same logic, sensitive content such as financial statements should always be handled by human translators. As a rule of thumb:
- Use raw machine translation for low-impact, quickly perishable, and unambiguous content. Think internal documentation, user-generated content like product reviews (for which consumers generally expect low quality), customer inquiries, or frequently amended content like feature and information updates.
- Apply light or full MTPE when your brand may suffer as a result of inaccuracies. It’s the case of product titles and descriptions (they tend to contain proper names and polysemous words), knowledge bases, press releases, etc.
- Leave it in the hands of human translators whenever top quality is key and there are no monetary or time constraints. Think legal, financial, or technical documents for which accuracy and consistency are paramount, as well as highly creative collateral designed to drive action.
Because machine translation is nowadays integrated in most modern CAT tools, you can aid the work of post-editors with CAT tools’ features such as glossaries, term bases, and translation memories, as well as brand books and style guides. This makes it very feasible to keep brand voice and key messaging consistent across cultures and languages with MTPE.
Translation management systems (TMS)
A translation management system (TMS) is a software platform designed for translation project management at scale.
It streamlines the translation process from start to finish, creating a workflow from the initial request to project completion, and allowing users to plan, assign, and manage translation projects in an organized manner.
TMS solutions vary according to the needs of users (e.g. some are cloud-based and some are on-site). Nevertheless, in general, they all include a core set of features:
- Support for multiple languages and file formats
- A dashboard that displays a project’s status and provides access to essential tools in one place
- The option to automatically push (import) content into the TMS for translation and pull (export) it back after project completion
- The ability to streamline processes by adding users or importing existing translation requests from other systems for easier collaboration with third-party providers
- The ability to assign, modify, or release tasks at any point in the project
- A reporting function that provides a clear overview of progress across all projects
- User role management that maintains a separation between those who manage projects and those who translate them
- Real-time collaboration for active teamwork and reduced project turnaround time
- CAT-tool functionalities such as glossaries, termbases, and translation memories
- Machine translation capabilities that help speed up the translation process
- Out-of-the-box integrations and application programming interfaces (API) for connectivity with existing systems and other tools such as content management systems (CMS), UI design tools, version control systems, and more
The best TMS solutions are flexible enough to allow users to create their own workflows and customize the software according to the needs of specific projects, departments, and teams. They are highly scalable and grow together with your business.
How to choose the right translation technology vendor
Because of the diversity of translation software out there, companies can pick and choose exactly what kinds of solutions work best for them from a solution provider. However, how do you tell whether a translation technology provider is the right fit for your organization?
Translation technology providers must not only offer the ideal mix of functionalities and tools; they need to be flexible enough to integrate with existing systems while supporting scalability. You also need to assess whether their technology is intuitive for your users.
Creating a translation technology request for proposal (RFP)will help you articulate the features and benefits of your ideal translation technology provider. Doing so will allow you to sift through the plethora of providers and point out which ones meet all of your requirements.
Before creating the RFP, some points to consider when you’re evaluating potential providers include:
- Look for providers that offer concrete examples of work they’ve done with other clients.
- Find a provider that can help you choose which features are best for your business. Translation technology providers should also be able to offer advice on how the tools will affect workflow and improve processes.
- Prioritize vendors that cover everything you need now (e.g., a must-have integration with one of your existing systems) but also has the flexibility to adapt with your business’ needs in the future.
- Consider what kind of support you’ll need. Translation technology vendors should offer training sessions and troubleshooting assistance whenever necessary.
- Decide whether you want an on-premises or cloud-based solution.
- Choose a provider that offers the most advanced AI-powered machine translation.
- Finally, picking a tool that can be used on any operating system (Linux, Mac, Windows) from anywhere is essential considering freelancers will work remotely.
The future of translation technology
Translation tools are already much more nuanced than they were even just a couple of decades ago thanks to machine learning. Translation technology will only grow stronger as artificial intelligence continues to advance.
However, translation will never be completely automated because human translators and specialized machine translation tools have their own advantages that complement each other well (e.g. humans are great at creatively thinking outside the box, and AI is ideal for automating repetitive tasks).
As a result, experts agree that the future of translation will combine human intelligence and AI-powered machine translation to produce the best results. In the end, human translators and AI tools will work together as a single translation team with each member playing its distinctive role. We will likely transition from computer-assisted human translation to human-assisted computer translation.
Machine-translation post-editing will ensure that AI-generated translations are accurate and fluent, while human translators will make the final creative changes to cater for specific audiences.
The horizon looks bright: Translation technology will continue allowing translators to focus on more complex translations, leaving the tedious and repetitive tasks for AI-powered tools.
It’s no longer a question of whether or not to use translation technology; the next question is: Which one?
Last updated on March 31, 2023.