S&AS: INT: Autonomous Experimentation Platform for Accelerating Manufacturing of Advanced Materials

发现和制造新材料是一个既费力又耗时的过程。从历史上看，这一过程需要材料科学家和制造工程师付出巨大的努力，因为他们探索了许多不同的成分组成和工艺条件，以找到正确的组合，从而获得具有所需性能的材料。该项目正在开发一个自主实验平台，以取代大量的人工实验工作，从而加快先进材料的发现和制造。这样一个自主的实验试验台可能会对工程实践产生重大影响，并彻底改变材料发现和先进制造的格局。这反过来将加速在大量行业和应用中发现材料。参与该项目的学生将接受数据科学和材料/制造科学相结合的培训。该项目将人工智能和自主模块引入自主实验平台，以模仿人类科学家处理令人惊讶的观察、合成不同的知识体系以及探索复杂而巨大的设计空间的能力。该平台的关键研究部分围绕三个能力主题组织：(1)开发以有效确定设计空间中最有前途的区域，(2)探索以识别不寻常设计引起的意外并对其进行推理，以及(3)基于从文献和数据库中挖掘新知识来扩展新发现的设计空间，优先获得可能包含优秀材料设计解决方案的区域的知识。拟议的系统具有认知性、适应性、丰富的知识和可执行的任务，通过简单的命令与人类科学家互动，并在最小和适当程度的人工干预下执行自主发现过程。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

Discovering and manufacturing new materials is a laborious and time-consuming process. Historically this process takes significant effort from material scientists and manufacturing engineers as they explore many different ingredient compositions and process conditions to find the right combination leading to a material with the desired properties. This project is developing an autonomous experimentation platform to replace a significant amount of this human effort in experimentation and thus accelerate the discovery and manufacturing of advanced materials. Such an autonomous experimental testbed could have significant impacts on engineering practice and revolutionize the material discovery and advanced manufacturing landscape. This would in turn accelerate discovery in materials for a great number of industries and applications. Students to be involved in the project will receive training with a blend of data science and material/manufacturing science. This project introduces artificial intelligence and autonomy modules into an autonomous experimentation platform to mimic a human scientist's ability to handle surprising observations, synthesize diverse bodies of knowledge, and explore a complex and large design space. The key research components in this platform are organized around three capabilities themes: (1) exploitation to efficiently determine the most promising regions of a design space, (2) exploration to recognize and reason about surprises arising from unusual designs, and (3) expansion of newly-discovered design space based on mining new knowledge from literature and databases, preferentially gaining knowledge in regions likely to contain superior material design solutions. The proposed system is cognizant, adaptive, knowledge-rich and taskable, interacting with human scientists by way of simple commands and executing an autonomous discovery process with a minimal and appropriate degree of human intervention.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.