CROTTC-IF Architecture Overview
Figure 1 : Overview of the CROTTC-IF architecture. From left to right, it comprises the CROTTC AM (blue), the lightweight IF-MDD LM (orange), and the detailed IF components.
Paper context
Paper title: Beyond Acoustic Sparsity and Linguistic Bias: A Prompt-Free Paradigm for Mispronunciation Detection and Diagnosis Abstract: Mispronunciation Detection and Diagnosis (MDD) requires modeling fine-grained acoustic deviations. However, current ASR-derived MDD systems often face inherent limitations. In particular, CTC-based models favor sequence-level alignments that neglect transient mispronunciation cues, while explicit canonical priors bias predictions toward intended targets. To address these bottlenecks, we propose a prompt-free framework decoupling acoustic fidelity from canonical guidance. First, we introduce CROTTC, an acoustic model enforcing monotonic, frame-level alignment to accurately capture pronunciation deviations. Second, we implicitly inject mispronunciation information via the IF strategy under the knowledge transfer principle. Experiments show CROTTC-IF achieves a 71.77% F1-score on L2-ARCTIC and 71.70% F1-score on the Iqra'Eval2 leaderboard. With empirical analysis, we demonstrate that decoupling acoustics from explicit priors provides highly robust MDD. Passages referencing this figure: stic fidelity. • State-of-the-Art Performance: Operating entirely without auxiliary data or explicit canonical prompts, the CROTTC-IF framework achieves highly competitive results. It demonstrates strong generalization across diverse benchmarks, ranging from general L2 English corpora (L2-ARCTIC, ERJ, speechocean762) to the specialized Arabic Qur'anic recitation task (Iqra'Eval2). 2 Related Works Figure 1 : Overview of the CROTTC-IF architecture. From left to right, it comprises the CROTTC AM (blue), the lightweight IF-MDD LM (orange), and the detailed IF components. Currently, modern MDD approaches broadly fall into two categories: dictation-style and text-prompting style. 2.1 Dictation-Style MDD and the Acoustic Trap Dictation-style methods aim to recognize the uttered phoneme sequence e or shadowing scenarios [ mine22_SP , geng2024APSIPA , geng2025perception ] , restricting the real-world applicability of MDD. 3 Consistency Regularization on Optimal Temporal Transport Classification In this section, we analyze the inherent limitations of CTC and introduce the architecture of our proposed frame-wise dense acoustic model: CROTTC . Our approach, as illustrated in the left panel of Figure 1 , is built upon two core pillars: Consistency Regularization (CR) and Optimal Temporal Transport Classification (OTTC) . 3.1 Limitations of Connectionist Temporal Classification Connectionist Temporal Classification (CTC) [ graves2006connectionist ] has established itself as a fundamental criterion for sequence-to-sequence modeling, particularly in the ASR domain. By introducing a blank t ly models the frame-to-label correspondence rather than marginalizing over sparse paths. 3.3 Consistency Regularization While OTTC captures fine-grained mispronunciation cues, acoustic models can be
