解决神经算子在水下声传播预测中的频谱偏差问题
Mitigating Spectral Bias in Neural Operators for Underwater Transmission Loss Prediction
- 分阶段处理：先全局粗略预测，再局部精细修正
- 在南海数据集上精度显著优于FNO，推理速度保持毫秒级
- 适合需要高精度声场建模的海洋声学应用
快速准确预测水下声学传播损耗对实时海洋声学应用至关重要。尽管傅里叶神经算子（FNO）因具有全局感受野而成为强大的代理模型，但其存在频谱偏差问题：FNO中的频率截断机制会滤除高频成分，导致预测结果过度平滑，无法捕捉细微的干涉模式。为克服这一局限，本文提出谱-空间残差学习（S2RL）框架。S2RL将预测任务分解为由粗到精的过程：首先通过谱域全局传播器生成全局一致的预测，再由空间域局部修正器恢复高频残差。在南海数据集上的实验表明，该方法显著优于FNO基线，同时保持毫秒级推理速度。
原文摘要 · Abstract (English)
Predicting underwater acoustic transmission loss rapidly and accurately is crucial for real-time ocean acoustic applications. While Fourier Neural Operators (FNO) have emerged as powerful surrogate models due to their global receptive fields, they suffer from spectral bias. The frequency truncation mechanism in FNO filters out high-frequency components, resulting in over-smoothed predictions that fail to capture fine-grained interference patterns. To overcome this limitation, this paper proposes a Spectral-Spatial Residual Learning (S2RL) framework. S2RL decomposes the prediction task into a coarse-to-fine process: a spectral Global Propagator first generates a globally consistent prediction, and a spatial Local Refiner subsequently recovers the high-frequency residuals. Experimental results on a South China Sea dataset show that the proposed method significantly outperforms FNO baselines while maintaining millisecond-level inference speeds.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。