Machine learning models to extract accurate material transport

学生将开发机器学习模型，从高保真模拟以及高功率激光实验获得的时间序列数据中提取准确的材料传输和强度系数。这些将用于提高我们对极端条件下材料的理解，特别是在存在涉及湍流和磁化等离子体的非局部过程的情况下。该项目的目标是开发一种新的图形神经网络框架，以解决材料特性的复杂微观物理问题，并使用符号回归技术通过封闭的数学表达式来识别它们的涌现行为。这具有实际应用，从航天器组件的设计到星系团能量平衡的建模。

The student will develop machine learning models to extract accurate material transport and strength coefficients from time-series data, obtained from high-fidelity simulations as well as experiments with high-power lasers. These will be used to improve our understanding of materials under extreme conditions, particularly in presence of non-local processes involving turbulent and magnetized plasmas. The goal of the project is to develop a novel graph neural network framework to address the complex micro-physics of material properties and to identify their emergent behaviour via closed mathematical expressions using symbolic regression techniques. This has practical applications ranging from the design of spacecraft components to the modelling of the energy balance in cluster of galaxies.