This repository contains all the project I have done in Graph Neural Network Domain.
The entity alingmnet project is my Summer Internship project of 2021 in IISC Bangalore.
Where I worked under the supervision of Dr. Sundeep Prabhakar Chepuri, on solving entity alignment problems. The knowledge graph is a graph that is represented as a tuple of three values [V, E, R], it has several relational edges making it very informative. Given two knowledge graph, entity alignment identifies entities in each graph which are similar and bring them closer together. Entity alignment is useful when we want to club knowledge graphs together such that the same item/node is not repeated. It becomes one of the most important preprocessing steps before using big graphs for several other purposes.
I developed a method to solve the entity alignment using a graph neural network. My motivation for the architecture was siamese network You can find in dept details and discussion in my summer internship report.
A simple but effective network, run over the DBP15K dataset