8月25日
11:23
11:23官方账号arXiv cs.LG@Xiaoyang Xie, Clarence W. Rowley
This paper introduces the Inertial Manifold Neural Operator (IMNO) for solving dissipative PDEs, offering better interpretability, accuracy, and stability. IMNO-SE, a variant for shift-equivariant PDEs, preserves symmetry and improves performance. Benchmark experiments demonstrate IMNO's effectiveness.
推荐理由:这篇论文提出了IMNO,一个解决耗散PDEs的新方法,比传统神经网络操作符更稳定准确,特别适合对耗散系统的研究。