用少量数据实现印地语语法纠错，仅靠指令微调和确定性解码。
Minimal-Edit Instruction Tuning for Low-Resource Indic GEC
- 基于指令微调与轻量约束解码，避免数据增强
- 马拉雅拉姆语得分为92.41，印地语81.44，排名前列
- 适合资源稀缺语言的低资源语法纠错场景
印地语族语言的语法纠错面临标注数据稀缺、文字多样和形态复杂的问题。本文提出无需数据增强的方法，使用指令微调的大语言模型结合保守解码策略。采用120亿参数的GEMMA 3模型，在bnb 4-bit精度下通过参数高效微调（PEFT）与Alpaca风格格式进行指令微调。解码采用确定性、约束感知的流程，辅以轻量级归一化器，确保修改最小且语义不变。推理阶段通过从训练数据中构建的确定性错误分类器的分类体系、标签分布与优先级排序，直接生成特定语言的固定提示。在官方未微调的GLEU评测中，马拉雅拉姆语得分92.41（第六），印地语81.44（第三）。结果表明，分类器引导的提示设计、适配器式指令微调与确定性解码为印地语族语法纠错提供了一种可复现且计算高效的替代方案，也启发未来对更强形态句法约束与人类中心的保守编辑评估研究。
原文摘要 · Abstract (English)
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.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。