The Indian Institutes of Technology have launched an AI-driven open-source system that translates scientific and technical content from English to Hindi and other Indian languages, as the apex body for technical education, AICTE, has opened the doors to engineering education in eight Indian languages.
To boost the study of engineering in regional languages, the team aims to translate 500 engineering texts into Hindi within one year and in 15 Indian languages in three years. Using artificial intelligence, the system is expected to finish the task at one-sixth the speed at which translators would work manually.
Project Udaan, as the initiative is known, was seeded seven years back when Professor Ganesh Ramakrishnan, institute chair professor, department of computer science and engineering, IIT Bombay, and his team analysed the wide gap in availability of technical knowledge in Hindi and other Indian languages.
Commenting on the deployment of machine-based translation for the project, Krishnaswamy VijayRaghavan, principal scientific adviser to the government of India, said: “Because of technology, because we can create glossaries in every area, because we can scan and digitise and translate, for most kinds of documents you can get a high level of accuracy from machine-based translation.”
He, however, pointed out that “there will still be a huge requirement for human intervention which will highlight the nuances that are essential to translation.”
He said the effort speaks to the deep understanding of society that is required to excel. “To excel in a field, you must have rootedness in your culture, in your society and use the power of your understanding which comes from your language. If you look at China, Japan, Sweden or Germany, education is primarily in their native language but bi-lingual or trilingual capabilities are common.”
Professor Ramakrishnan said, “Our approach to machine translation has been that it will be aided by human effort. We started building lexicons of various technical domains. We have developed robust bilingual OCR technology and several post-editing tools by which we now have access to digital bilingual dictionaries in machine readable format. We are, therefore, able to use the appropriate scientific and technical terms available in Hindi instead of transliterating the English terms.”
By deploying an AI-based translation engine, a technical book can be translated in less than one-sixth the time it would take for a team of domain and linguistic experts working manually. “In due course, we expect to achieve a much shorter turnaround time,” Ramakrishnan said.
(This is a slightly modified version of an article originally published in Times of India. The original article can be found at https://timesofindia.indiatimes.com/india/artificial-intelligence-breaks-language-barrier-for-tech-edu/articleshow/86218960.cms)