One email each morning — yesterday's AI, sortedGet it in your inbox

METAL LAB

NE-BERT: A Multilingual Language Model for Nine Northeast Indian Languages

arXiv:2608.180942026-08-20

Large pretrained language models have demonstrated remarkable capabilities across diverse languages, yet critically underrepresented low-resource languages remain marginalized. We present NE-BERT, a domain-specific multilingual encoder model trained on approximately 8.3 million sentences spanning 9 Northeast Indian languages and 2 anchor languages (Hindi, English), a linguistically diverse region with minimal representation in existing multilingual

Authors · Badal Nyalang

Read on arXiv

Latest papers

All papers →

Latest from METAL LAB