Speaker Bio

Dawei Wang is a professor at the School of Microelectronics, Xi'an Jiaotong University. He received his B.S. and M.S. degrees in Physics from Peking University in 2000 and 2002, respectively. He received his Ph.D. from Queen's University (Canada) in 2008. He did postdoctoral work with Professor Laurent Bellaiche (APS fellow) at the University of Arkansas. He has presided or co-presided over several projects supported by the National Natural Science Foundation of China. His research focuses on the theoretical and numerical simulation of perovskites and the strategy of applying artificial intelligence and machine learning to simulation. He has discovered important physical phenomena, such as Fermi Resonance and Fano Resonance in perovskites. His publications include about 60 research papers in prestigious high-level international journals, including Nature Materials, Physical Review Letters, Nature Communications, Physical Review B, Advanced Functional Materials, and so on.

王大威，西安交通大学微电子学院教授。2000, 2002年获得北京大学物理学学士与硕士学位。2008年在加拿大女王大学获得博士学位。之后在学Laurent Bellaiche（APS fellow）教授组（阿肯色大学）进行了博士后工作。主持或参与国家自然科学基金面上、重点等项目多项。研究工作集中于钙钛矿结构材料的理论与计模拟研究，以及将人工智能、机器学习应用于模拟的策略。发现了钙钛矿中包括费米共振和Fano共振在内的重要物理现象，发表60余篇研究论文，包括Nature Materials，Physical Review Letters，Nature Communications， Physical Review B, Advanced Functional Materials等。

Abstract

Modeling is merely a human construct - not the real world - but it is indispensable to our research and helps us better understand the real complex world. Traditionally, people use three techniques to overcome various difficulties in research: simplified models, approximate solutions, and developing multiscale models. As machine learning exemplified by AlphaGo and AlphaGo Zero increasing importance in recent years, machine-learning-assisted modeling and physics-informed machine learning are gaining traction. I introduce significant recent advances in modeling and machine learning and relate them to our research on perovskites and other systems.

虽然模型只是人造的工具而不是现实世界，但其对于理解复杂的真实世界是必不可少的。传统上在对真实世界的探索中，人们主要使用三种技术以克服研究中的种种困难：简化模型、数值近似解以及发多尺度模型。近年来随着以AlphaGo和AlphaGo Zero所代表的机器学习的兴起，机器学习辅助建模和物理知识机器学习正日益受到重视。在报告中，我将介绍建模和机器学习的一些最新重要进展，并展示建模与机器学习在我们对钙钛矿和其他体系的研究中的应用。