Banner
Workflow
Navbar

IIT Bombay launches Project Udaan

IIT Bombay launches Project Udaan

  • If one speaks or writes in a particular language, but others can’t understand it. A barrier in communication is created, hampering the flow of messages.
  • Keeping this in mind, IIT Bombay launched Project Udaan, an AI-based translation ecosystem to translate engineering textbooks and learning materials in Hindi.
  • In due course, textbooks in all domains can be taken up.

Project Udaan

  • ‘Project Udaan,’ a donation-based project, is an end-to-end ecosystem, which translates scientific and technical content from English to Hindi and all other Indian languages.
  • The AI-based translation software ecosystem can translate engineering textbooks, learning materials in one-sixth of the time taken by a team of domain and linguistic experts working manually on it.
  • The project is also in line with the New Education Policy, which promotes imparting of higher education through Indian languages.
  • Further, the Project aims to translate 500 engineering texts in Hindi in one year and in 15 Indian languages in 3 years.
  • In addition to this, among the various goals of the Project, important one includes:
  1. Enabling a web platform with a complete translation workbench, domain-specific linguistic resources for efficient and consistent quality translation
  2. Set up a parallel corpus in the technical domain (>30,00,000 words) to speed up the translation and post-editing research.

Similar Projects

  • The institute has developed similar projects that focus on Indian languages, such as Akshar Anveshini, MALTA, DECILE, INDICOCR, and more.
  • Akshar Anveshini is an optical character recognition (OCR) and post-editing system for the Sanskrit language.
  • It converts the document images into an editable electronic format that enables data compression, search and edit options.
  • MALTA or Multi-Modal and Multi-Lingual Temporal Sentence Alignment localizes sentences/ captions in videos that leverage both audio and video modalities and that can generalize to new and possibly low-resource language settings.

Categories