Natural language translation with graph convolutional neural network

• Main Author: Zhu, Yimin
• Other Authors: Xavier Bresson
• Format: Final Year Project
• Language: English
• Published: 2018
• Subjects: DRNTU::Engineering::Computer science and engineering::Computing methodologies::Artificial intelligence; DRNTU::Engineering::Computer science and engineering::Computing methodologies::Document and text processing
• Online Access: http://hdl.handle.net/10356/73963
• Institution: Nanyang Technological University

With the trend of artificial intelligence, scientists and researchers developed dozens of methods to use AI in different aspects of our daily life. Natural language processing is one of the most popular areas using AI. When deal with natural language, AI scientists always use recurrent neural network(RNN) to train the AI since it fits the structure of sentences naturally. However, RNN is not suitable if we want to fully utilize the computation resource of hardware. Researchers at Facebook AI Research group come up with the idea to use convolutional neural network(CNN) to release the whole power of GPUs.This project aims to realize a neural network for language translation that uses graph convolution technique(GCNN) instead of traditional CNN in order to improve the performance of accuracy and training speed. The experiments and results are discussed in details.