Collaborative Research: CyberTraining: Implementation: Medium: The Informatics Skunkworks Program for Undergraduate Research at the Interface of Data Science and Materials Science

该项目将开发一种可持续和可扩展的方法，以培养在材料信息学中机器学习（ML）的研究和应用方面熟练的劳动力。材料信息学中的机器学习正在以前所未有的能力迅速改变材料科学与工程（MS&amp；E），扩展材料数据库，改进材料模拟，挖掘文本，自动化材料研究和开发，加速材料设计。正如包括美国国家科学院和美国矿物、金属和材料协会（TMS）在内的多项近期研究所指出的那样，为材料信息学培训下一代机器学习劳动力，以实现其通过先进材料改善人类状况的巨大潜力，这一点至关重要。不幸的是，材料信息学中的机器学习几乎完全没有出现在本科阶段的材料课程中。然而，信息学工具的力量与它们在科学与工程领域的快速发展和相对新颖性相结合，为本科生（UGs）通过有影响力的研究积极学习创造了机会。基于这一动机，该项目将创建新的基础设施和生态系统，供美国各地的UGs参与和培训材料信息学中的应用机器学习研究，称为informatics Skunkworks。Skunkworks由从事材料信息学研究的导师/UG团队组成。该项目为团队提供了新的资源，包括课程、软件和研究问题，以及一个支持研究和有效协作的实践社区。机器学习和材料信息学的低成本和可访问性为Skunkworks吸引研究资源有限的导师和学生创造了机会，特别是在服务于代表性不足群体的机构中。Skunkworks是一种可持续和可扩展的方法，可以通过培训在材料信息学的机器学习研究和应用方面熟练的多样化劳动力来满足这一未满足的需求。该项目将免费提供(a)培训UG相关ML、材料信息学和研究专业发展的课程，(b)增强现有ML包的软件工具，使UG可以访问，以及(c)真实和适当级别的研究问题。拟议的工作还将建立一个实践社区，使富有成效的导师/UG研究团队网络能够有效和协作地使用项目开发的课程、软件和其他资源，以支持转变未来的劳动力。拟议工作的智力价值是(a)开发可扩展的资源，以增加UG在数据科学和材料科学与工程边界研究中的经验和学习，(b)发展一个从事材料信息学研究的导师和UG研究人员社区，以及(c)通过劳动力发展和材料信息学培训增加数据科学工具的利用，以解决ms&e和相关领域的关键问题。拟议工作的更广泛影响是：(a)自由传播材料信息学的课程和工具，(b)在广泛适用的研究和专业技能方面培训员工，导师和UGs，以及(c)为材料信息学研究人员建立一个多样化的实践社区。该项目将通过主要支持通常研究资源有限的UG机构和社区学院，使能够开发先进材料信息学的新劳动力，特别是对于代表性不足的群体。该项目由计算机和信息科学与工程理事会的高级网络基础设施办公室资助，数学和物理科学理事会的材料研究部也提供资金。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

The project will develop a sustainable and scalable approach to train a workforce skilled in research and application of machine learning (ML) in materials informatics. ML in materials informatics is rapidly transforming materials science and engineering (MS&E) by an unprecedented ability to extend materials databases, improve materials simulation, mine texts, automate materials research and development, and accelerate materials design. As pointed out by multiple recent studies, including from the National Academies and the Minerals, Metals & Materials Society (TMS), it is essential to train a next generation workforce in ML for materials informatics to realize its enormous potential for improving the human condition through advanced materials. Unfortunately, ML in materials informatics is almost completely absent from today's materials curricula at the undergraduate level. However, the power of informatics tools combined with their rapid evolution and relative novelty in MS&E creates an opportunity for engaging undergraduates (UGs) with active learning through impactful research. With this motivation, the project will create new infrastructure and an ecosystem for the engagement and training of UGs across the U.S. in research using applied ML in materials informatics, called the Informatics Skunkworks. The Skunkworks consists of mentor/UG teams performing research in materials informatics. The project provides the teams with new resources, consisting of curricula, software, and research problems, and with a community of practice to support research and to work effectively and collaboratively. The low cost and accessibility of ML and materials informatics creates an opportunity for Skunkworks to engage mentors and students with limited research resources, particularly at institutions serving underrepresented groups. The Skunkworks is a sustainable and scalable approach that can fulfill this unmet need by training a diverse workforce skilled in research and application of ML for materials informatics. The project will provide freely available (a) curriculum to train UGs in relevant ML, materials informatics and research professional development, (b) software tools that augment existing ML packages to be UG accessible, and (c) authentic and appropriate-level research problems. The proposed work will also develop a community of practice to enable a network of productive mentor/UG research teams to effectively and collaboratively use the curricular, software and other resources developed by the project to support transforming the future workforce. The intellectual merit of the proposed work is to (a) develop scalable resources to increase UG experience and learning in research at the boundary of data science and materials science and engineering, (b) grow a community of mentors and UG researchers engaged in materials informatics research, and (c) increase the utilization of data science tools for solving critical problems in MS&E and related fields through workforce development and materials informatics training. The broader impact of the proposed work is to (a) freely disseminate enabling curricula and tools for materials informatics, (b) train staff, mentors and UGs in broadly applicable research and professional skills, and (c) develop a diverse community of practice for materials informatics researchers. The project will enable the development of a new workforce capable of advanced materials informatics, especially for underrepresented groups, by supporting primarily UG institutions and community colleges that often have limited research resources. This project is funded by the Office of Advanced Cyberinfrastructure in the Directorate for Computer and Information Science and Engineering, with the Division of Materials Research in the Directorate for Mathematical and Physical Sciences also contributing funds.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.