Sequential Inputs For Neural Networks

Question

What type of neural networks should be used when we have sequential inputs. Let say we want to phrase some text which should give output depend on the sequence of the words. For this the output should dependent on the previous state of the inputs.

Answer 1

Recurrent neural networks are what you are looking for -- at least in the general case, where sequences are long, or not all sequences have the same length.

Here's an example, here is a Python implementation and here are some reference papers.