Apologies, but no results were found for the requested archive. Perhaps searching will help find a related post.
The use of technology in the translation industry is now standard. As such the use and application of Artificial Intelligence (AI) in translations continues to grow.
Nevertheless, how useful they are depends on the task to be completed. When we think about AI and translation, machine translation is what comes to mind. We have previously discussed machine translation to increase productivity in certain instances. However, automation in translation doesn’t necessarily mean machine translation.
Automated translations can be described as the act of automating the translation process by the use of technology. The application that provides the automated translations is installed on a server and is accessed in a similar way to a server-based translation memory. In other words, the end user or content manager simply “pushes” the translation by clicking a button, deciding whether to use machine, human or hybrid translation processes and then the translation provider pushes back the translations completed.
There are other technologies that can help us divide texts into categories that machines can easily read, without the use of a project manager. That does not mean human translators will disappear from translations any time soon, but automation can help streamline some parts of the translation process. For example, integrating your Content Management System (CMS) with a Translation Management System (TMS) can help you automate website translation, by eliminating the process of preparing and transferring the files among translators.
Going back to the automation from a pure translation perspective, we have to admit that now there is a true revolution taking place. Machine translation has been around for decades but it was based on statistical models that fed on huge amounts of data. What is different now is the use of large neural networks to predict the chances of having a sequence of words, shaping entire sentences in a single model. This has created a higher level of accuracy in output from neural machine translation compared to those from statistical machine translation. In practice, companies that are heavy translation users are taking some leaps into implementing this technology on their workflow, with human supervision… or not, at their own risk. There are many instances, like with creative type of content or uncommon languages, where neural machine translation is used only if you are looking for a good laugh.
At Idea Translations we work to find the most suitable solution to our client’s translation problem. We create an optimal translation workflow by means of a thorough evaluation of the content and of the key performance indicators to define the return on investment of their translation initiative, and then develop a workflow that could include Translation Automation on a number of fronts without compromising quality.