Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This study investigates the effectiveness of a homogeneous ensemble of monolingual models, specifically five independently fine-tuned GujaratiBERTs, in Named Entity Recognition (NER) for the Gujarati language, which faces unique challenges due to its morphology and syntax. By comparing this approach to a baseline and several heterogeneous models, the research finds that the language-aligned ensemble outperforms all alternatives, achieving an F1 score of 0.8442. These results highlight the potential of leveraging monolingual pretraining over complex architectures in low-resource language tasks.
Language-aligned ensembles can outperform complex architectures in low-resource NER tasks, achieving a significant F1 score boost in Gujarati.
Named Entity Recognition (NER) for Gujarati remains underexplored, hindered by the absence of capitalization cues, rich morphology, lexical ambiguity, and free word order. Prior ensemble work has emphasized architectural diversity by combining heterogeneous classifiers, multilingual encoders, or classical sequence models, rather than exploiting language-aligned monolingual pretraining. This study asks whether, for a low-resource, morphologically rich language like Gujarati, a homogeneous ensemble of a single monolingual encoder outperforms such architectural diversity. We propose HomoEnsNER, a homogeneous ensemble of five independently fine-tuned GujaratiBERT models combined via majority voting, evaluated against a single GujaratiBERT baseline and six heterogeneous alternatives, including combinations with MuRIL-base, MuRIL-large, IndicBERT, mBERT, BiLSTM, CRF, and a stacked BiLSTM-CRF-GujaratiBERT architecture. All eight models were trained under a consistent budget and evaluated using entity-level F1 on the Naamapadam Gujarati test split. HomoEnsNER achieved the highest F1 (0.8442), surpassing the baseline (0.8347) and every heterogeneous alternative (lowest: 0.7855), indicating that language alignment is a more effective, budget-conscious ensembling strategy than architectural complexity for low-resource Indian language NER.