Datraw, Advancing Collaborative Data Collection, and Recognition Models for Low Resource Languages in the Research Community

Abstract

There is a significant need for a platform for all that promotes collaborative data collecting and supports community-driven advances in task recognition in the current era of increasing low resource data-driven research. “Datraw,” a groundbreaking platform meant to connect researchers with the larger research community, serves as a centralized hub, empowering researchers to upload sample data with labels, while inviting linguists, common individuals, and fellow researchers to contribute by drawing data, whether in the form of characters or complete works and build rich and expansive dataset for low resource language and offers the opportunity to leverage the amassed datasets for training basic classification models. Through comprehensive validation and review methods, the platform assures data integrity and quality, ensuring that the obtained data stays correct and relevant for recognition tasks.