Beyond Native Scripts: Robust Language Identification for Romanized and Code-Mixed Indian Languages
Date:
Delivered an invited presentation covering robust Language Identification (LID) strategies under script variation.
- Presented core processing challenges regarding script loss, linguistic clustering shifts, and systemic data alignment confusion in multilingual Indian contexts.
- Outlined a structural curriculum-based training strategy alongside critical evaluation insights for optimization.