What does machine translation mean?
The text is translated by a computer. There are different methods of machine translation.
For rule-based methods, the designated language is linked to dictionaries. The same goes for syntactic and morphological rules. In simple terms, the following happens: the MT engine searches the content to be translated in its dictionary, adapts it to the grammar rules of the target language and generates the translation. This sounds simple. However, since language often doesn’t adhere to its own rules, there are many exceptions.
- Statistical MT
This method uses large bilingual text corpora to pair source and target language with statistical methods based on their frequency and proximity. The system derives translation rules according to which the text is then translated. The translated texts usually come closer to natural language than rule-based translated texts, and the translation usually also works very well with frequently occurring language pairs. For this, however, very large amounts of data are required. Languages with fewer or smaller text corpora are difficult to translate with this method.
- Neural networks
A completely different approach to machine translation is provided by neural networks or methods of "deep learning". In this case, different nodes called neurons, in accordance with the human nervous system, take over individual parts of the main task and learn independently on the basis of previously defined characteristics.
Can I let the computer translate my texts?
In principle, yes. Machine translation is often worthwhile for large quantities of text which must be translated cost-effectively in the shortest possible time.
Contact us. We can advise you if and to what extent a machine translation is suitable, and help to choose the right system for you.
I already have an MT system. Can I use it with ONTRAM?
This is not a problem. We integrate your MT system seamlessly into ONTRAM, and you benefit from the combination of the latest language technologies to make your translation more efficient and increase quality.