It is worth remembering that seq2seq can be used for more than just Machine Translation. Let's see some examples:
- Lukasz Kaiser, in Grammar as a Foreign Language (https://arxiv.org/abs/1412.7449), uses the seq2seq model to build a constituency parser. A constituency parse tree breaks text into subphrases. Non-terminals in the tree are types of phrases, the terminals are the words in the sentence, and the edges are unlabeled.
- Another application of seq2seq is SyntaxNet, aka Parsey McParserFace, (a syntactic parser; https://research.googleblog.com/2016/05/announcing-syntaxnet-worlds-most.html), which is a key first component in many NLU systems. Given a sentence as input, it tags each word with a part-of-speech (POS) tag that ...