Minimal-Edit Instruction Tuning for Low-Resource Indic GEC
This addresses the problem of grammatical error correction for low-resource Indic languages, offering a computationally efficient alternative to augmentation-based methods, though it appears incremental in its technical approach.
The paper tackles grammatical error correction for Indic languages with limited supervision by proposing an augmentation-free approach using instruction-tuned large language models and conservative decoding, achieving GLEU scores of 92.41 on Malayalam (sixth overall) and 81.44 on Hindi (third overall).
Grammatical error correction for Indic languages faces limited supervision, diverse scripts, and rich morphology. We propose an augmentation-free setup that uses instruction-tuned large language models and conservative decoding. A 12B GEMMA 3 model is instruction-tuned in bnb 4-bit precision with parameter-efficient fine-tuning (PEFT) and Alpaca-style formatting. Decoding follows a deterministic, constraint-aware procedure with a lightweight normaliser that encourages minimal, meaning-preserving edits. We operationalise inference, subsequent to instruction fine-tuning (IFT), via a fixed, language-specific prompt directly synthesised from a deterministic error classifier's taxonomy, label distributions, and precedence ordering computed on the training corpus. Under the official untuned GLEU evaluation, the system scores 92.41 on Malayalam, sixth overall, and 81.44 on Hindi, third overall. These results indicate that classifier-informed prompt design, adapter-based instruction tuning, and deterministic decoding provide a reproducible and a computationally efficient alternative to augmentation-centred pipelines for Indic GEC. The approach also motivates future work on stronger morphosyntactic constraints and human-centred evaluation of conservative edits.