At the Google For India 2018 conference in New Delhi yesterday Google launched its AI platform Navlekha, which enables publishers to make offline Indian language content fully editable and streamlines the online publishing process.
Navlekha uses machine learning technology to convert Indian language PDFs and scanned files into editable text to help publishers share content online. “Navlekha” derives from the Sanskrit for “a new way to write.”
India is home to the second-largest population of Internet users in the world, and adds 40 million each year. But many Indian publishers lack the technologies and channels to publish their content on the Internet. Google estimates that 90 percent of Indian publishers don’t have a content distribution website, and that the amount of Indian-language content now on the Internet is but one percent of the English-language total.
Google also announced a new spoken website feature in Google Go that uses NLP and speech synthesis to allow users to “listen to websites,” even on 2G connections. The feature supports 28 languages, including Hindi, Bengali, Malayalam, Marathi and Tamil. Google has also integrated Hindi-friendly functions into flagship products Google Search and Google Go, and users in India can now customize their Google Search results to include both English and Hindi sources.
Google clearly sees the huge potential for development of local language content and services on the subcontinent, and plans to use machine learning to tap into this expanding market.
Author: Alex Chen | Editor: Michael Sarazen