III: Small: Collaborative Research: Robust Materials Genome Data Mining Framework for Prediction and Guidance of Nanoparticle Synthesis

该提案支持新的计算机科学研究，为新技术提供基础。 本论文的研究目标是设计新的稳健的数据挖掘和机器学习算法，以解决复杂材料基因组数据挖掘中的计算挑战。美国政府发起的材料基因组计划研究旨在快速、低成本地发现、制造和部署先进材料，这为解决清洁能源、国家安全和人类福利方面的挑战提供了巨大的机会。然而，由于前所未有的规模和复杂性，主要的计算挑战是全面材料基因组数据分析的瓶颈。迫切需要新的数据挖掘和机器学习策略来弥合差距并促进新材料的发现。为了解决挖掘这种全面的异质材料基因组数据的关键和挑战性问题，PI建议开发一种新的强大的数据挖掘，并探索如何整合来自多个数据源的特征。PI将在线提供已开发的计算方法和工具，供公众使用。这些方法和工具有望影响其他材料基因组和生物化学研究，并使研究人员能够有效地测试新材料设计的性能预测假设。所提出的算法和工具，预计将有助于知识提取的应用程序在更广泛的科学领域与大量的高维和异构的数据集。这个项目将促进开发新的教育工具，以加强目前的几门课程。PI建议开发一个新的强大的数据挖掘框架，旨在探索以下研究任务。首先，PI将开发新的计算工具来自动化材料基因组数据处理，包括通过新的鲁棒秩-k矩阵完成方法进行缺失值填补，基于鲁棒张量因子分解的特征提取方法，以及使用鲁棒主动学习模型进行信息纳米颗粒选择。其次，PI将研究新的稀疏多任务多视图学习模型，以整合异质材料表征，用于预测催化能力和理论建模测量的关联。第三，为了预测新合成纳米颗粒的催化能力，PI将通过研究弹性嵌入，自适应损失，L1范数图和有向图模型来设计新的鲁棒半监督学习模型。 提出的稀疏多视图特征学习和鲁棒半监督学习模型满足了大规模数据分析和集成的关键需求。这种独特的能力将使新的计算应用在大量的研究领域。它的进步，从而扩大了工程创新和计算分析之间的关系。

This proposal supports novel computer science research to provide foundations for new technologies. The research objective of this proposal is to design new robust data mining and machine learning algorithms for solving the computational challenges in complex materials genome data mining. The Materials Genome Initiative research has been launched by U.S. government to discover, manufacture, and deploy advanced materials fast and low-cost, which holds great opportunities to address the challenges in clean energy, national security, and human welfare. However, the major computational challenges are the bottlenecks for comprehensive materials genome data analysis due to unprecedented scale and complexity. There is a critical need for new data mining and machine learning strategies to bridge the gap and facilitate the new materials discovery. To solve the key and challenging problems in mining such comprehensive heterogeneous materials genome data, the PIs propose to develop a novel robust data mining and explore ways to integrate features from multiple data sources. The PIs will make the developed computational methods and tools online, available to the public. These methods and tools are expected to impact other material genome and biochemistry research and enable investigators working on new material design to effectively test performance prediction hypothesis. The proposed algorithms and tools are expected to help knowledge extraction for applications in broader scientific domains with massive high-dimensional and heterogonous data sets. This project will facilitate the development of novel educational tools to enhance several current courses.The PIs propose to develop a novel robust data mining framework targeting to explore the following research tasks. First, the PIs will develop new computational tools to automate the material genome data processing, including missing values imputation by a new robust rank-k matrix completion method, robust tensor factorization based feature extraction approach, and informative nanoparticles selection using robust active learning model. Second, the PIs will investigate the new sparse multi-task multi-view learning model to integrate heterogeneous material characterizations for predicting the catalytic capabilities and associations to theoretical modeling measurements. Third, to predict the catalytic capabilities of the new synthesized nanoparticles, the PIs will design novel robust semi-supervised learning models by investigating elastic embedding, adaptive loss, L1-norm graph, and directed graph models. The proposed sparse multi-view feature learning and robust semi-supervised learning models meet the critical needs of large-scale data analysis and integration. Such unique capabilities will enable new computational applications in a large number of research areas. It advances and thus extends the relationship between engineering innovation and computational analysis.