SBIR Phase II: Accelerating R&D through Streamlined Machine Learning Algorithms for Small Data Applications in Advanced Manufacturing

该小型企业创新研究（SBIR）第二阶段项目的更广泛/商业影响将改善和加速制药行业新化学品，工艺和配方的开发。通过利用机器学习（ML）的力量，该项目旨在节省高达95%的时间和资源，同时减少废物的产生，从而增强药物开发的商业和社会影响。药物开发中的传统过程优化是一项耗时且昂贵的奋进，通常依赖于试错法。临床前药物开发中的先导化合物优化和路线搜索可能需要数月时间，涉及数千次实验，仅人员费用就高达数百万美元。该项目旨在通过使用ML来指导小数据集的实验设计来解决这些挑战。通过ML和化学知识的结合，该项目旨在通过仅建议最有希望的实验并最大限度地减少失败尝试的数量来简化优化过程。这一解决方案不仅赠款患者更快地获得药物，还能让企业更早地获得收入，并保持更长的市场独占期。这一小型企业创新研究二期项目解决了研发化学家面临的最重大挑战之一：优化合成工艺中的分类变量，特别是溶剂选择。溶剂在化学工业中起着至关重要的作用，包括反应，分离，纯化和配制。合适的溶剂可以提高效率，降低成本，并产生更环保的工艺。尽管在ML引导的优化和绿色溶剂选择方法方面取得了成功的进展，但可用的工具不能有效地将环境和性能参数联合收割机组合以同时进行溶剂选择和工艺优化。该项目将为化工行业的科学家提供利用小数据和创新机器学习技术推进制造的解决方案。该奖项反映了NSF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/commercial impact of this Small Business Innovation Research (SBIR) Phase II project will improve and accelerate the development of new chemicals, processes, and formulations in the pharmaceutical industry. By harnessing the power of machine learning (ML), this project aims to save time and resources by up to 95%, while reducing waste generation, thereby enhancing the commercial and societal impact of drug development. Traditional process optimization in drug development is a time-consuming and expensive endeavor, often relying on trial-and-error approaches. Lead optimization and route scouting in pre-clinical drug development can take months and involve thousands of experiments, costing millions of dollars in personnel expenses alone. This project seeks to address these challenges by employing ML to guide experimental design with small datasets. Through a combination of ML and chemistry knowledge, this project aims to streamline the optimization process by suggesting only the most promising experiments and minimizing the number of failed attempts. This solution not only grants patients quicker access to medicines but also enables companies to generate earlier revenue and maintain longer market exclusivity.This Small Business Innovation Research Phase II project addresses one of the most significant challenges faced by research and development (R&D) chemists: the optimization of categorical variables in synthetic processes, specifically solvent selection. Solvents play a vital role in the chemical industry, including reaction, separation, purification, and formulation. The proper solvent can improve efficiency, reduce costs, and result in a more environmentally friendly process. Despite successful advances in ML-guided optimization and green solvent selection methodologies, available tools do not effectively combine environmental and performance parameters for simultaneous solvent selection and process optimization. This project will provide a solution for scientists across the chemical industry to leverage small data and innovative ML technologies to advance manufacturing.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.