DigSAL Project Overview

Addressing Linguistic Diversity and Digital Inequality

South Asia, home to over 2 billion people, boasts a rich linguistic tapestry with dozens of major languages and hundreds of minority languages. These languages primarily belong to two major language families: Indo-European and Dravidian. Despite their deep historical interactions, both groups remain underexplored in linguistic and computational research compared to European, Semitic, and East Asian languages. This lack of study stems from a shortage of trained linguists specializing in South Asian languages, limited computational resources, and a general lack of understanding between linguists and computer scientists. While advanced Language Models (LLMs) like GPT-3 and GPT-4 have revolutionized Natural Language Processing (NLP) for languages such as English, South Asian languages lag significantly behind. The scarcity of large datasets and the computational power required for modern NLP techniques, including neural networks and Deep Learning, exacerbate this gap. Technologies like Google Translate and voice assistants (e.g., Alexa, Siri) have made strides in many languages, but South Asian languages, especially minority ones, remain under-resourced. This disparity creates a significant Digital Divide, leaving millions without access to essential digital services and information available to speakers of well-resourced languages.

Our Proposal: Reducing Inequality through Innovation

Our proposal aims to address this Digital Divide, aligning with the Sustainable Development Goal (SDG) 10: Reduced Inequality. To achieve this, we will also focus on SDG 9: Industry, Innovation, and Infrastructure, and SDG 17: Partnerships for the Goals. Our collaborative efforts will involve building infrastructure and developing innovative technologies to create necessary resources and computational tools. This will empower industries to develop robust NLP and AI applications tailored for South Asian languages.

Building on a Strong Foundation

We will leverage the knowledge and experience gained from previous collaborative projects between the University of Konstanz (Germany), the University of Moratuwa (Sri Lanka), and the University of Engineering and Technology (Pakistan). These initiatives have laid the groundwork for computational linguistic applications and resources for indigenous languages in Sri Lanka and Pakistan. Our focus will include:

1. Training and Education: Developing educational materials for Natural Language Processing and Computational Linguistics, specifically geared towards South Asian languages, to train both linguists and computer scientists.

2. Resource Development: Enhancing digital text, speech, and image (script) processing capabilities for South Asian languages, including research into language structure (morphology, syntax, semantics).

3. NLP Applications: Creating tools such as speech recognizers, optical character recognizers, and translators, including advanced Natural Language Understanding (NLU) systems.

4. Infrastructure Enhancement: Ensuring that universities in South Asia have access to the necessary infrastructure to support state-of-the-art Deep Learning technology and intensive NLP computing.

To achieve our goals, we plan to extend our existing collaborations to include technical and administrative staff from our partner universities. This will help establish sustainable infrastructure for large-scale NLP computing, ensuring the long-term success of our initiative.

Join us in bridging the Digital Divide in South Asia and creating a more inclusive digital future. Together, we can harness the power of technology to reduce inequality and foster innovation across linguistic boundaries!