0
$\begingroup$

I am doing a university report and it seems that encode-decode RNNs are optimal for machine translation. I would like something else to compare it to, but I can't seem to make a proper Google search for it.

Can you recommend what else can be used for machine translation that I can compare RNNs to?

$\endgroup$
1
  • $\begingroup$ Please consider upvoting the answers if you found them useful, and marking one as correct if deemed so. Alternatively, please comment on what you think the answers are lacking. $\endgroup$ – noe Feb 18 at 12:07
0
$\begingroup$

Recursive Neural Networks assume you have a relation between the current input and the previous input, if you organize data correctly CNNs can be used in place, some CNNs models have score higher than RNNs in BLEU metric. There is an article in Medium about it Why Convolutional Neural Networks are a Great Architecture for Machine Translation

Some implementation examples are Facebook's architecture proposed in the paper A novel approach to neural machine translation and this one from a Stanford's professor.

Also CNN's have been used by Google for language understanding and even in some fast-compact models or their new AI that talks on the phone.

$\endgroup$
0
$\begingroup$

Transformers.

Most NMT systems since 2018 use the Transformer architecture.

All relevant NMT frameworks support Transformers, including:

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.