How to create a neural network for text transformation?

Question

Machine learning

Please tell me, can someone come across this, there is a large database of text pairs, for example: it was "gfsjfshkll" -> it became "hhlshlghslkhgklhs" the algorithm is not clear. Is it possible to train a neural network so that it understands the algorithm and transforms the text based on the list of these pairs. In neural networks, not even a beginner, rather interested. Searches have shown that you need to move towards recurrent neural networks and learning with a teacher. Perhaps someone solved similar problems, please direct where to move, perhaps there are ready-made solutions for an example. And one more question, how productively can a neural network predict a transformation or a correspondence of this kind?

Answer 1

short: maybe.
longer: you need to have a large dataset, because this is a classic training with a teacher