Searching for Next Generation Spintronic Materials Using Machine Learning

使用机器学习寻找下一代自旋电子材料

• 批准号: 523137-2018
• 负责人: Burgess, Jacob
• 金额: $ 1.82万
• 依托单位: University of Manitoba
• 依托单位国家: 加拿大
• 项目类别: Engage Grants Program
• 财政年份: 2018
• 资助国家: 加拿大
• 起止时间: 2018-01-01 至 2019-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=646542
• 关键词: Searching; Next; Generation; Spintronic; Materials

Engineering of materials is an increasingly important aspect of modern technology. Tuning physical material**properties offers incredible flexibility not only to optimize performance and efficiency of devices such as**computer transistors, solar cells, sensors, and many others, but also an avenue to realize fundamentally new**device concepts such as novel quantum qubits. Materials science remains extraordinarily challenging simply**due to the huge range of combinations of atomic composition, structure, and synthesis techniques that can be**applied. Guiding research is incredibly challenging. Lumiant Corporation is a Canadian company developing a**machine learning based artificial intelligence, Xaedra, that can be taught to predict material structures and**compositions that will yield desirable properties. To achieve this, Xaedra must be fed a large amount of**verified experimental data and theoretical calculations relating to any particular physical property. The**proposed project aims to explore Xaedra's applicability to predicting dynamic properties in magnetic materials.**Specifically, the project will teach Xaedra how to predict a property called damping which determines how**much energy is dissipated during a magnetodynamic change. Minimizing damping is crucial in the**actualization of practical 'spintronics' - a computing scheme that aims to use magnetism to replace electronic**charge to create a new breed of ultra-efficient computers. Success in this project will significantly expand**Xaedra's capabilities in predicting magnetic properties and dynamic properties. It will also demonstrate**Xaedra's value as a tool in a very active and commercially prominent area of materials science research.

材料工程是现代技术日益重要的一个方面。调整物理材料**属性提供了令人难以置信的灵活性，不仅可以优化**计算机晶体管、太阳能电池、传感器等设备的性能和效率，而且还提供了实现全新**设备概念（例如新型量子）的途径量子位。由于可以应用的原子组成、结构和合成技术的组合范围广泛，材料科学仍然极具挑战性。指导研究非常具有挑战性。 Lumiant Corporation 是一家加拿大公司，开发基于机器学习的人工智能 Xaedra，可以学习预测材料结构和成分，从而产生所需的性能。为了实现这一目标，Xaedra 必须获得大量经过验证的实验数据和与任何特定物理特性相关的理论计算。 **拟议项目旨在探索 Xaedra 在预测磁性材料动态特性方面的适用性。**具体来说，该项目将教 Xaedra 如何预测称为阻尼的特性，该特性决定磁动力变化期间耗散的能量**量。最小化阻尼对于实现实用的“自旋电子学”至关重要，“自旋电子学”是一种计算方案，旨在利用磁性取代电子电荷来创建新一代超高效计算机。该项目的成功将显着扩展**Xaedra 预测磁特性和动态特性的能力。它还将展示**Xaedra 作为一个非常活跃且商业突出的材料科学研究领域的工具的价值。