|
And the translator then has less work to do with proofreading, which saves further time and money. Ultimately, the quality of your translation should be as close as possible to a professional human translation. And the better the translation machine is trained, the easier it will be for the translator to achieve good quality during post-editing. Conversely, this means that the quality of the initial machine translation for difficult texts and language combinations as well as a small training corpus can be so poor that the translator is more inclined to translate from scratch.
Otherwise, the risk of overlooking mistranslations or omissions is simply too high. The following HK Phone Number decision always needs to be made: Is the machine translation good enough to use without post-editing in some cases? For example, for purely informational purposes. Or should the quality be as close as possible to that of a human translation and the text should therefore be post-edited by a translator. If this is the case, then the initial machine translation must be so good that the translator really needs much less time for post-editing. Otherwise the use of MT makes no sense. Ian:And finally, who is considering machine translation is asking themselves: “What about the costs, how much can I save.
Tim:At the beginning I talked about the immense computing power that we have available to us today. This has made handling large amounts of data much cheaper. However, training the translation machine is still a cost that should not be ignored. So it depends on how much content you have that the translation engine can learn about. The more content, the more expensive it initially becomes. On the other hand, you also save in terms of post-processing time and costs. So if you invest in training the translation engine in advance using existing translations and terminology, you can save time and money down the line.
|
|