Predicting Critical Transitions in Multiscale Dynamical Systems Using Reservoir Computing

Predicting the properties of molecular materials: multiscale simulation workflows meet machine learning

Machine Learning tools are nowadays widely applied extensively to the prediction of the properties of molecular materials, using datasets extracted from high-throughput computational models. In several cases of scientific and technological relevance, the properties of molecular materials are related to the link between molecular structure and phenomena occurring across a wide set of spatial scales, from the nanoscale to the macroscale. Here, we describe an approach for predicting the properties of molecular aggregates based on multiscale simulations and machine learning.

Operator learning for predicting multiscale bubble growth dynamics

Simulating and predicting multiscale problems that couple multiple physics and dynamics across many orders of spatiotemporal scales is a great challenge that has not been investigated systematically by deep neural networks (DNNs). Herein, we develop a framework based on operator regression, the so-called deep operator network (DeepONet), with the long term objective to simplify multiscale modeling by avoiding the fragile and time-consuming "hand-shaking" interface algorithms for stitching together heterogeneous descriptions of multiscale phenomena. To this end, as a first step, we investigate …

两相储层流问题的多尺度摄动方法

The Multiscale Perturbation Method for Two-Phase Reservoir Flow Problems

本文提出两相流多尺度摄动方法（MPM-2P），结合区域分解与多尺度Robin耦合方法，通过重用基函数避免完整更新，在保证精度的同时将两相流模拟计算成本降低近90%。

解码神经元网络：一种用于预测连接性和功能性的储备池计算方法

Decoding Neuronal Networks: A Reservoir Computing Approach for Predicting Connectivity and Functionality

针对神经元网络电生理测量分析的挑战，本研究提出基于储备池计算网络（RCN）的模型，可解码时空数据并在宏观尺度重建网络结构，其连接图预测性能优于互相关等常用方法，且经实验验证能预测网络对局部光遗传学刺激等特定输入的响应。

基于多维混合量子储备池计算的低维湍流预测

Forecasting Low-Dimensional Turbulence via Multi-Dimensional Hybrid Quantum Reservoir Computing

针对复杂动力学预测难题，提出结合量子演化与经典记忆增强的混合量子-经典储备池架构，在两类混沌流体模型中实现了更优的预测性能，为非线性时间序列建模提供新平台。

参数感知储备池计算机的动力学分析

Dynamical analysis of a parameter-aware reservoir computer

本研究针对参数感知储备池计算预测动力系统临界转变的机制问题，结合动力系统理论与数值方法，揭示训练后储备池映射的Neimark-Sacker分岔特性，为相关机器学习算法的运作提供理论洞见。