abumafrim/Hausa-Visual-Genome-Dataset

### Barka da zuwa!!! --> Welcome!!! # Hausa Visual Genome: A Dataset for Multi-Modal English to Hausa Machine Translation ## Welcome! This is a repository for Hausa Visual Genome dataset, containing about 33K images and a short English and Hausa description of a section of each image. Jump straight to one of the sections below, or just scroll down to find out more. ## Table of Contents - The Problem - Language Resources Developed - Dataset Formats - Data Statistics - Using The Dataset - Paper(s) from this Project - The Team - Contact Us - Glossary ## The Problem Multi-modal Machine Translation (MMT) enables the use of visual information to enhance the quality of translations. The visual information can serve as a valuable piece of context information to decrease the ambiguity of input sentences. Despite the increasing popularity of such a technique, good and sizeable datasets are scarce, limiting the full extent of their potential. Hausa, a Chadic language, is a member of the Afro-Asiatic language family. It is estimated that about 100 to 150 million people speak the language, with more than 80 million indigenous speakers. This is more than any of the other Chadic languages. Despite a large number of speakers, the Hausa language is considered a low-resource language in natural language processing (NLP). This is due to the absence of sufficient resources to implement most of the tasks in NLP. While some datasets exist, they are either scarce, machine-generated, or in the religious domain. Therefore, there is the need to create training and evaluation data for implementing machine learning tasks and bridging the research gap in the language. ## Language Resources Developed Hausa Visual Genome is a multimodal dataset consisting of text and images suitable for English-to-Hausa multimodal machine translation task and multimodal research. We selected all the English segments (captions) of the Hindi Visual Genome along with associated images and au …