A Fusion-Aware Two-Stage Framework for Mispronunciation Detection and Diagnosis in Low-Resource Modern Standard Arabic
This work addresses the problem of accurate phoneme recognition for mispronunciation detection in low-resource Arabic, which is important for computer-assisted language learning.
The paper presents a two-stage end-to-end framework for mispronunciation detection and diagnosis in low-resource Modern Standard Arabic, achieving an F1-score of 0.7201, a 63.1% relative improvement over the baseline, and ranking first in the IqraEval.2 Challenge.
Accurate phoneme recognition is pivotal for mispronunciation detection and diagnosis (MDD) in modern standard Arabic (MSA), yet remains constrained by data scarcity and the synthetic-real domain gap. This work proposes a two-stage end-to-end framework. It integrates a pre-trained encoder with causal dilated temporal convolutional networks to preserve fine-grained phonetic variations. A hierarchical two-stage strategy first learns general mappings from native/synthetic corpora, then adapts to scarce real learner data to mitigate domain shift without over-correction. Prediction stability is further enhanced via multi-checkpoint ensemble inference with N-gram rescoring. Evaluated on the QuranMB.v2 test set, our system achieves an F1-score of $0.7201$, a $63.1$\% relative improvement over baseline ($0.4414$). This performance ranks at the top of the IqraEval.2 Challenge, establishing a new state-of-the-art for low-resource MSA in MDD.