arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2407.13010cs.LGcs.CEphysics.comp-phstat.ML
收藏

独立于分辨率的神经算子

A Resolution Independent Neural Operator

Hopkins Extreme Materials Institute(霍普金斯极端材料研究所);Dept. of Civil and Systems Engineering(土木与系统工程系);Johns Hopkins University(约翰·霍普金斯大学);Dept. of Chemical and Biomolecular Engineering(化学与生物分子工程系);Dept. of Applied Mathematics and Statistics(应用数学与统计学系)

Bahador Bahmani, Somdatta Goswami, Ioannis G. Kevrekidis, Michael D. Shields

AI总结:

本文提出了一种独立于分辨率的神经算子架构RINO,通过自适应学习连续基函数,实现对任意采样输入输出函数的高效处理。

摘要:

深度算子网络(DeepONet)是一种强大的神经算子架构,它使用两个神经网络来在无限维函数空间之间进行映射。这种架构允许在域内的任何位置评估解场,但要求输入函数在相同位置进行离散化,限制了实际应用。我们介绍了一个从输入输出数据中学习算子的一般框架,允许任意传感器位置和数量。这首先引入了一个独立于分辨率的DeepONet(RI-DeepONet),它能够处理任意但足够精细离散化的输入函数。为此,我们提出了两种字典学习算法,能够自适应地从相关信号上任意点云中学习连续基函数,这些基函数参数化为隐式神经表示(INRs)。这些基函数将输入函数数据投影到有限维嵌入空间,使其与DeepONet兼容而无需改变架构。我们特别使用正弦表示网络(SIRENs)作为可训练的INR基函数。同样,字典学习算法识别输出数据的基函数,定义了新的神经算子架构:独立于分辨率的神经算子(RINO)。在RINO中,算子学习任务简化为将输入基函数的系数映射到输出基函数。我们通过多个数值示例展示了RINO在处理任意采样的输入和输出函数时的鲁棒性和适用性。

英文摘要:

The Deep Operator Network (DeepONet) is a powerful neural operator architecture that uses two neural networks to map between infinite-dimensional function spaces. This architecture allows for the evaluation of the solution field at any location within the domain but requires input functions to be discretized at identical locations, limiting practical applications. We introduce a general framework for operator learning from input-output data with arbitrary sensor locations and counts. This begins by introducing a resolution-independent DeepONet (RI-DeepONet), which handles input functions discretized arbitrarily but sufficiently finely. To achieve this, we propose two dictionary learning algorithms that adaptively learn continuous basis functions, parameterized as implicit neural representations (INRs), from correlated signals on arbitrary point clouds. These basis functions project input function data onto a finite-dimensional embedding space, making it compatible with DeepONet without architectural changes. We specifically use sinusoidal representation networks (SIRENs) as trainable INR basis functions. Similarly, the dictionary learning algorithms identify basis functions for output data, defining a new neural operator architecture: the Resolution Independent Neural Operator (RINO). In RINO, the operator learning task reduces to mapping coefficients of input basis functions to output basis functions. We demonstrate RINO's robustness and applicability in handling arbitrarily sampled input and output functions during both training and inference through several numerical examples.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑