Skip to content
Conference Open access

Indic Machine Translation: A Survey with Challenges and Upcoming Directions

Sep 2026 · 2026 IEEE International Conference on Advances in Computing Research On Science Engineering and Technology (ACROSET) · pp. 1-6 · 0 citations · 31 references

Abstract

Machine Translation has become much more sophisticated than rule-based systems and is now represented by neural architectures, but still has pitfalls to face even when it comes to Indic languages spoken by more than 1.5 billion people on Earth. This general review paper analyzes how machine translation models have evolved starting with Rule Based Machine Translation (RBMT), then Statistical Machine Translation (SMT), to Neural Machine Translation (NMT), with emphasis on the Indic language issues. We discuss the advantages and drawbacks of each paradigm, outline some of the main milestones, and continue challenges such as resource intensive language-supporting, domain-specific, cultural washes, and ethical issues. The survey lists such emerging trends as Large Language Models (LLMs), multimodal translation, the ability to code-switch, as well as human-in-the-loop workflows. Our discussion not only offers comprehensive information on what Indic machine translation can currently achieve, but also where it can go in the future, where it is important to note that linguistically inclusive translation systems that are culturally sensitive are necessary.

Read PDF

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.