Developing a Hybrid Neural Network for Part-Of-Speech Tagging and Named Entity Recognition

Abstract

Enabling our computers to understand human languages is very important as we move towards the 4th to 5th industrial revolution. A lot of efforts are already made to fast track this development most especially for highly resourced languages such as English, French, German, among others. The most notable open-source NLP tool built is the Natural Language toolkit, a collection of modules and corpora that provides researchers in Natural Language Processing (NLP) with extremely useful tools and resources. In the Philippines, several researchers have contributed to this advancement mostly for the Filipino language. In this paper, we introduce the design architecture of a hybrid neural network model that combines the best component features of the existing architectures for the POS Tagging and NER tasks. We also present initial experiment results.