米佐语自动语音识别（ASR）语音语料库：结合形态感知评估的Whisper与SraVaani 1.0微调
A Speech Corpus for Mizo Automatic Speech Recognition: Whisper and SraVaani 1.0 Fine-Tuning with Morphology-Aware Evaluation
AI总结：
本研究构建米佐语语音语料库，通过微调Whisper与SraVaani 1.0模型开发米佐语ASR系统，发现Whisper-large-v3结合形态感知评估WER仅7.22%，SraVaani 1.0经米佐语微调后性能显著提升。
AI中文摘要：
本研究针对低资源语言米佐语开发自动语音识别（ASR）系统，流程包括收集17.62小时语音数据、整理数据，并使用三款Whisper多语言模型及SraVaani 1.0印度多语言模型对米佐语ASR系统进行微调。Whisper-large-v3的常规词错误率（WER）最低，为18.08%；而形态感知评估得到的WER为7.22%。对SraVaani 1.0印度多语言模型进行零样本评估时，其WER为58.27%；针对米佐语的微调将其常规WER降至29.45%，形态感知WER降至17.93%。结果表明，Whisper模型即使适配未接触过的语言也能实现极低WER；相比之下，SraVaani 1.0在其多语言模型中支持米佐语，但使用精心整理的米佐语语音数据进行微调可大幅提升其性能。
英文摘要：
This study reports the development of an Automatic Speech Recognition (ASR) system in Mizo, a low-resource language. The development included collecting 17.62 hours of speech data, curating it, and fine-tuning the Mizo ASR system with three Whisper multilingual models and with the SraVaani 1.0 Indic multilingual model. Whisper-large-v3 achieved the lowest conventional WER (18.08%), while morphology-aware evaluation yielded a WER of 7.22%. Zero-shot evaluation of the SraVaani 1.0 Indic multilingual model yielded a WER of 58.27%, while Mizo-specific fine-tuning reduced the conventional WER to 29.45% and the morphology-aware WER to 17.93%. The results demonstrate that the Whisper model can achieve a substantially low WER, even when adapted to an unseen language. In contrast, SraVaani 1.0 supports the Mizo language in its multilingual model; however, fine-tuning with carefully curated Mizo speech data substantially improves its performance.
