SuvarthiSarkar/Creation-of-Knowledge-Graphs-for-Low-resource-Language

In this paper, we have worked on two different research problems: performing cross-lingual Knowledge Graph (KG) alignment and performing entity-alignment between English and Hindi parallel corpora. We have analyzed different architectures performing cross-lingual KG alignment. Of all the models, graph convolutional networks(GCNs) are pretty intuitive due to their simple architecture and efficient output.We were able to implement crosslingual entity alignment successfully for english to chinese translation. Our goal was to do the same for a low resource Indian language (Hindi) but we were not able to find such entity aligned corpora for low resource Indian language and thus worked on making our own entity alignment from english to hindi by using Neural Machine Translation and Statistical Machine translational models like IBM Model 1 and Hidden Markov Model.