Reproduce “X-MDD: LLM-Based Mispronunciation Detection and Diagnosis for Unseen Languages with Canonical References” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Xintong Wang, Junchuan Zhao, Ye Wang
X-MDD tests whether Mandarin- and English-trained mispronunciation detection transfers to Arabic Qur’anic recitation. It gives the model learner speech, intended phones, synthesized canonical speech, and labeled Arabic demonstrations at inference, then trains it to detect errors and recognize actual phones. It reaches 38.07% detection F1, below an Arabic-trained baseline’s 44.14%; evaluation covers one unseen language.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.