machine learning - Sequence to sequence modeling in python -


i trying make chatbot uses sequence sequence model respond user's input. problem input sequence given model never same. input sequence list of words. have created vocabulary maps each word in sequence own unique id, however, input still variable , not fixed can't use sequence sequence model. understand possible use encoder map sequence of words fixed vector representation , have decoder map vector sequence.

the question have how go encoding sequence of words fixed vector? there sort of technique used purpose?

mapping sequence of words vector representation can accomplished recurrent neural network. can take @ introduction: http://colah.github.io/posts/2015-08-understanding-lstms/

there tutorial in tensorflow tool kit address sequence sequence mapping architecture example code: https://www.tensorflow.org/versions/r0.11/tutorials/index.html

before working rnn, however, recommend going through basics neural networks: http://deeplearning.net/software/theano/tutorial/#basics

bengio's deep learning book: http://www.deeplearningbook.org/ covers lot of materials rnn, involves quite bit of math.


Comments

Popular posts from this blog

sql server - Cannot query correctly (MSSQL - PHP - JSON) -

php - trouble displaying mysqli database results in correct order -

C++ Linked List -