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

Researchers introduced NE-BERT, a multilingual encoder model trained on 8.3 million sentences across 9 Northeast Indian languages plus Hindi and English, which outperforms IndicBERT-V2 and MuRIL on all 9 languages with 15.97x and 7.64x lower average perplexity, respectively, and 1.50x better tokenization fertility than mBERT. The model, released under CC-BY-4.0, addresses vocabulary fragmentation in low-resource languages like Pnar and Kokborok through aggressive upsampling, and downstream part-of-speech tagging validates its practical utility.

arXiv:2608.18094v1 Announce Type: new Abstract: 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 models. By employing weighted data sampling and a custom SentencePiece Unigram tokenizer, NE-BERT outperforms IndicBERT-V2 and MuRIL across all 9 Northeast Indian languages, achieving 15.97X and 7.64X lower average perplexity respectively, with 1.50X better tokenization fertility than mBERT. We address critical vocabulary fragmentation issues in extremely low-resource languages such as Pnar 1,002 sentences and Kokborok 2,463 sentences through aggressive upsampling strategies. Downstream evaluation on part-of-speech tagging validates practical utility on three Northeast Indian languages. We release NE-BERT, test sets, and training corpus under CC-BY-4.0 to support NLP research and digital inclusion for Northeast Indian communities.