You can try this yourself on Google Translate

inoccotepola222

However, if the syntax is very different, such as E.g. in the English-German language pair, usually more work needs to be done. Even with more creative texts, machine translation is often wrong. But to get back to the question “Has machine translation gotten better?”, yes, definitely. Machine translation of simple sentences is often surprisingly good. And the reason for this is that very similar or identical content has already been translated by a human translator. This is exactly what you have to keep in mind: machine translation always happens on the basis of a human translation.


The magic word is machine learning, which means that the machine recognizes the patterns HK Phone Number of texts that have already been translated (by humans), learns from them and can apply them to new texts. Ian:So are we basically talking about a huge database that compares new texts with existing data and can apply it? Tim:Of course it's a bit more complex. Using algorithms, the machine can analyze and evaluate the entire content, recognize patterns and learn from them. Of course, this doesn't happen in a single pass. At the beginning you train the machines with the existing data, i.e. translations and glossaries. The machine is then retrained at regular intervals with new translations to improve the results of the first machine translation.



The machine detects what was wrong and what was right through the changes made during post-editing by a translator and learns from them. Entering company- or context-specific terminology also increases quality. For example, if you are looking for the German translation for an English term, there are often several options. In a specific context, however, only one translation could be correct. It is therefore incredibly important that the machine also understands the context. The more data the machine can evaluate and learn from, the better the decisions it makes regarding syntax and terminology.