Advancing Natural Language Processing for Dialects and Linguistic Diversity

This project addresses challenges in natural language processing for dialects and linguistic diversity, with particular attention to low-resourced languages and varieties. We develop computational models that can effectively capture and process dialectal variation, focusing on innovative representation learning techniques and transfer methods for low-resource scenarios. Our approach combines advanced computational methods with linguistic insights to address four key challenges:

Our research outcomes will substantially improve dialect-resilient machine translation systems, enhance the ability of large language models to handle linguistic variations, and advance various NLP applications for low-resourced languages and dialects.