Table of Contents
Natural language processing is highly related to theoretical linguistics. This review aims to provide an overview of state-of-the-arts in the field of NLP under the categories in linguistics, through which I want to point out the meanings of our work.
This work is always going to be continued, and I will update it with the expansion of my knowledge in this field. It may perform incomplete at start. I will try my best to complete it!
Tasks of Natural Language Processing
The fields that are generally considered the core of theoretical linguistics are phonology, phonetics, morphology, syntax, and semantics.
Neural Networks for Morphology
(Minh-Thang Luong et al, 2013) employs recursive neural network(RNN) and neural language model(NLM) to construct word vectors for unseen complex words. Two models are proposed. One is called context-insensitive morphological RNN(cimRNN). Another is called context-sensitive morphological RNN(csmRNN).