Asiatic Society uses AI to decipher ancient Indian manuscripts

The Asiatic Society in Kolkata is utilizing AI transcription and machine learning to decode ancient manuscripts, making them accessible to global scholars, Arab News reported. Founded in 1784, it is one of India’s oldest research institutions, focusing on history, culture, and languages. With over 52,000 rare manuscripts, the society launched the Vidhvanika project in December to digitize and develop language models for ancient scripts.

Administrator Anant Sinha highlighted the need for expertise in various languages and scripts, with a team of scientists and specialists involved. Supported by the Center for Development of Advanced Computing, the project addresses a wide range of subjects in languages like Sanskrit, Arabic, and Bengali.

Decoding requires knowledge of scripts, language, and historical context, with few specialized scholars available. Sinha noted the current model accuracy is about 40%, aiming for 90-95% as the machine learning process evolves. The project was initiated on the 225th anniversary of James Prinsep’s birth, who deciphered ancient scripts crucial for understanding the Mauryan Empire’s history. Vidhvanika aims to preserve endangered languages and scripts, emphasizing the importance of understanding ancestral knowledge.