Multicon W 2023-24
Translate resource material and other texts from English to other Indian regional languages.
As part of a collaborative research initiative, I worked on the development of a cutting-edge AI-powered translation tool aimed at translating resource materials and other texts from English into various Indian regional languages. The project addressed the need to overcome language barriers in India, a country with a vast linguistic diversity, for sectors such as education, healthcare, e-governance, and regional content dissemination.
The tool leverages Neural Machine Translation (NMT), a modern AI technique that uses deep learning models to translate languages efficiently and accurately. Our implementation focused on using the MarianMT model, a transformer-based architecture known for its exceptional performance in machine translation. The MarianTokenizer was employed to preprocess and tokenize the input text, preparing it for translation. This approach ensures that the translations maintain the cultural and contextual relevance of the target language.