Enhancing Neural Machine Translation with Direct Preference Optimization Using Human Feedback for Indonesian, Thai, and Malay Languages

This project aims to enhance Neural Machine Translation (NMT) systems for three Southeast Asian languages: Indonesian, Thai, and Malay. Current NMT systems face challenges in maintaining contextual accuracy and fluency, especially when translating between these languages and others. The project will introduce Direct Preference Optimization (DPO) using human feedback to improve translation quality in the target languages, focusing on applications in the education sector. This involves creating a collaborative platform for research, knowledge exchange, and building educational networks among Southeast Asian universities and global research institutions.

Funded through BMBWF / OeAD GmbH (ASEA-UNINET).