Spectral-Bias and Kernel-Task Alignment in Physically Informed Neural Networks
원문 초록
Physically informed neural networks (PINNs) are a promising emerging method for solving differential equations. As in many other deep learning approaches, the choice of PINN design and training protocol requires careful craftsmanship. Here, we suggest a comprehensive theoretical framework that sheds light on this important problem. Leveraging an equivalence between infinitely over-parameterized neural networks and Gaussian process regression (GPR), we derive an integro-differential equation that governs PINN prediction in the large data-set limit -- the neurally-informed equation. This equation augments the original one by a kernel term reflecting architecture choices and allows quantifying implicit bias induced by the network via a spectral decomposition of the source term in the original differential equation.
함께 살펴볼 논문
공통 태그를 근거로 연결했습니다. 실제 인용 관계나 논문의 우수성을 뜻하지 않습니다.
Fast ground penetrating radar dual-parameter full waveform inversion method accelerated by hybrid compilation of CUDA kernel function and PyTorch
Trustworthy Prediction with Gaussian Process Knowledge Scores
Overcoming Overfitting in Reinforcement Learning via Gaussian Process Diffusion Policy
Fast Derivative Valuation from Volatility Surfaces using Machine Learning
Efficient Implementation of Gaussian Process Regression Accelerated Saddle Point Searches with Application to Molecular Reactions
Ensuring Reproducibility in Generative AI Systems for General Use Cases: A Framework for Regression Testing and Open Datasets
토론
이 논문과 연결된 포럼 글이 아직 없습니다.