New Models for Electronic Structure Prediction and Analysis

电子结构预测和分析的新模型

• 批准号: RGPIN-2016-05755
• 负责人: Pearson, Jason
• 金额: $ 1.82万
• 依托单位: University of Prince Edward Island
• 依托单位国家: 加拿大
• 项目类别: Discovery Grants Program - Individual
• 财政年份: 2016
• 资助国家: 加拿大
• 起止时间: 2016-01-01 至 2017-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=615884
• 关键词: New; Models; Electronic; Structure; Prediction

Most of the worlds major technological and societal challenges are being addressed through research in chemistry, including health, climate change, energy, security, and food supply. With rapid advances in computational power over the last two decades, advanced research computing (ARC) has now become an indispensable research tool in the chemical sciences and many other fields. Research scientists routinely rely on computational simulations to complement, and often exceed, what can be achieved with experiments alone, securing ARC as a transformative modern technology with broad socioeconomic impact in Canada. Unfortunately however, modern computational techniques for modelling chemical systems scale very poorly with molecular size, leading to a practical limit on the scope of molecular species accessible by computational means. Consequently, if computational techniques are to be reliably applied to the ever-increasing scope of important problems in chemistry, radically new computational models are required. Additionally, as the prevalence of computational simulation in chemistry expands, there is a concurrent need for techniques to distill relevant insight and meaning from resultant data, particularly for predictions of detailed electronic structures in quantum chemistry. Recent work in our group has made significant strides toward providing innovative solutions to these important problems in chemistry. We have worked to establish a comprehensive, open, and interactive data management platform for the chemical sciences with capabilities for pattern recognition and complex queries. These new technologies for data-driven inquiry will allow researchers to distill new knowledge from a vast chemical data landscape, which promises to yield new industrial catalysts, advanced materials, medicines, and transformative new computational simulation techniques among countless other substantial applications. In particular, we will utilize the technology to design radically new computational models based on machine learning algorithms applied to large datasets. Additionally, we will develop electronic structure analysis tools that predict and interpret the distribution of electrons within chemical bonds and lone pairs, which affords unique qualitative and quantitative insight into how molecular species interact and react.

世界上大多数主要技术和社会挑战正在通过化学研究得到解决，包括健康、气候变化、能源、安全和粮食供应。随着过去二十年计算能力的快速进步，高级研究计算（ARC）现已成为化学科学和许多其他领域不可或缺的研究工具。研究科学家通常依靠计算模拟来补充甚至超越单独通过实验所能实现的目标，从而确保 ARC 作为一项变革性的现代技术，在加拿大具有广泛的社会经济影响。然而不幸的是，用于模拟化学系统的现代计算技术与分子大小的关系非常差，导致计算手段可访问的分子种类的范围受到实际限制。因此，如果要可靠地将计算技术应用于化学中不断扩大的重要问题范围，就需要全新的计算模型。此外，随着化学计算模拟的普及，同时需要从结果数据中提取相关见解和意义的技术，特别是对于量子化学中详细电子结构的预测。 我们小组最近的工作在为化学中的这些重要问题提供创新解决方案方面取得了重大进展。我们致力于为化学科学建立一个全面、开放、交互式的数据管理平台，具有模式识别和复杂查询的能力。这些数据驱动探究的新技术将使研究人员能够从庞大的化学数据领域中提取新知识，从而有望产生新的工业催化剂、先进材料、药物和变革性的新计算模拟技术以及无数其他实质性应用。特别是，我们将利用该技术基于应用于大型数据集的机器学习算法来设计全新的计算模型。此外，我们将开发电子结构分析工具，预测和解释化学键和孤对电子内电子的分布，从而为分子物种如何相互作用和反应提供独特的定性和定量见解。