Translation model Architecture for question answering?

Hi Jeremy ,

I had gone through couple of papers released in April 2017 on question generation, where in they were using encoder - decoder architecture along with attention mechanism to generate questions given an answer, This architecture is used in machine translation as well. So my guess here is “question” (input) could be thought of as one language and answer(output) can be thought of another language , is my understanding correct here? If yes I would be glad to go ahead and apply that using fastai. Please help !!!



Yes they should be able to be handled by very similar models. We’ll be doing translation next week.

Awesome thanks !!!