I-Corps: Using machine learning methods and polymer data to predict properties of new polymers and accelerate application-specific polymer design

I-Corps项目更广泛的影响和商业潜力是开发用于聚合物选择、设计和发现的软件工具。该技术的潜在客户包括汽车、化工、石油和天然气等行业中制造或使用聚合物的公司的聚合物化学家或材料设计师。这项技术的价值在于克服了传统的、费力的、昂贵的材料开发的试错方法。聚合物制造业每年的研发支出总额约为100亿美元。这项技术有可能每年节省约1亿美元。此外，产品设计工作流程的加速可以大大缩短产品的上市时间，因为产品必须在整个产品生命周期中经历不断的变化。I-Corps项目基于数据驱动的机器学习工具的开发，以加速特定应用的聚合物设计和开发。机器学习（ML）算法在底层数据库上进行训练，生成预测模型，可以1)对尚未合成的新聚合物的性质进行即时预测，2)对满足设计目标的新聚合物和现有聚合物提出建议。目标是为最终用户公司创建定制的特定应用的聚合物设计工具，以及一个通用的数据到模型产品线，该产品线可以获取内部专有的（遗留/历史）聚合物数据，为聚合物制造公司生成预测模型。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 broader impact/commercial potential of this I-Corps project is the development of a software tool for polymer selection, design and discovery. Potential customers of this technology include polymer chemists or materials designers in companies that manufacture or utilize polymers in industries such as auto, chemicals, and oil and gas. The value of this technology lies in overcoming the traditional, laborious and expensive trial-and-error approaches to materials development. The total R&D expenditure of the polymer manufacturing industry is about $10 billion per year. This technology has the potential to save an estimated savings of $100 million per year. In addition, the acceleration of product design workflows could dramatically shorten time-to-market for products that must undergo evolutionary changes throughout the product’s life-cycle.This I-Corps project is based on the development of a data-driven machine learning tool to achieve accelerated application-specific polymer design and development. Machine learning (ML) algorithms trained on an underlying database produce predictive models, which can 1) make instantaneous predictions of properties of a new yet-to-be-synthesized polymer, and 2) make recommendations of new and existing polymers that will meet design objectives. The goal is to create a customized application-specific polymer design tool for end-user companies, and a generic data-to-model product line that ingests in-house proprietary (legacy/historic) polymer data to produce predictive models for polymer manufacturing companies.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.