Wei-Qiang Zhang

Low-velocity impact on carbon fiber reinforced polymer (CFRP) laminates induces barely visible internal damage that critically degrades structural integrity and poses significant safety risks if left undetected. Conventional finite element methods suffer from high computational costs and limited generalizability, while purely data-driven deep learning approaches lack physical interpretability and are prone to overfitting under small-sample conditions. To address these challenges, we propose a PINN framework that integrates multi-source heterogeneous information and physical constraints for the prediction of CAI and internal damage area in CFRP laminates. The model constructs a multimodal input system comprising surface damage profiles and process parameters. Extensive 5-fold cross-validation demonstrates that the proposed model achieves CAI strength prediction with R² = 0.8614 ± 0.0371 and RMSE = 36.96 ± 5.17 MPa, and damage area prediction with R² = 0.8329 ± 0.0228 and RMSE = 67.43 10.47 mm², confirming that mechanical information and internal damage details can be reliably inferred from surface profile data alone. Ablation and generalization studies further reveal that the embedded physical constraints effectively reduce training stochasticity and enhance model robustness and generalizability under limited data. Moreover, the predicted equivalent damage exhibits strong correlation with ultrasonic C-scan-derived damage area (Pearson's r > 0.91), validating the physical interpretability of the model. With its low parameter count and computational footprint, the proposed framework offers a promising pathway toward edge-deployable, real-time structural health monitoring of composite structures.

Machine anomalous sound detection (ASD) is a valuable technique across various applications. However, its generalization performance is often limited due to challenges in data collection and the complexity of acoustic environments. Inspired by the success of large pre-trained models in numerous fields, this paper introduces a robust ASD model that leverages self-supervised pre-trained models trained on large-scale speech and audio datasets. Although there are inconsistencies between the pre-training datasets and the ASD task, our findings indicate that pre-training still provides substantial benefits for ASD. To mitigate overfitting and retain learned knowledge when fine-tuning with limited data, we explore Fully-Connected Low-Rank Adaptation (LoRA) as an alternative to full fine-tuning. Additionally, we propose a Machine-aware Group Adapter module, which enables the model to capture differences between various machines within a unified framework, thereby enhancing the generalization performance of ASD systems. To address the challenge of missing attribute labels, we design a novel objective function that dynamically clusters unattributed data using vector quantization and optimizes through a dual-level contrastive learning loss. The proposed methods are evaluated on all benchmark datasets, including the DCASE 2020-2024 five ASD challenges, and the experimental results show significant improvements of our new approach and demonstrate the effectiveness of our proposed strategies.

This report introduces Dolphin, a large-scale multilingual automatic speech recognition (ASR) model that extends the Whisper architecture to support a wider range of languages. Our approach integrates in-house proprietary and open-source datasets to refine and optimize Dolphin's performance. The model is specifically designed to achieve notable recognition accuracy for 40 Eastern languages across East Asia, South Asia, Southeast Asia, and the Middle East, while also supporting 22 Chinese dialects. Experimental evaluations show that Dolphin significantly outperforms current state-of-the-art open-source models across various languages. To promote reproducibility and community-driven innovation, we are making our trained models and inference source code publicly available.