Autonomous chemical science and engineering enabled by self-driving laboratories
Autonomous chemical science and engineering enabled by self-driving laboratories
复制标题
DOI：
10.1016/j.coche.2022.100831
复制
发表时间：
2022-05-16
影响因子：
6.6
通讯作者：
Abolhasani, Milad
中科院分区：
文献类型：
--
作者：
Bennett, Jeffrey A.;Abolhasani, Milad
机器学习（ML）和人工智能的最新进展为通过自动驾驶实验室实现复杂反应系统的基础和应用研究的计算机化提供了令人兴奋的机会。自主机器人实验可以实现时间、材料和资源效率的探索和/或高维空间反应系统的优化。此外，对实验数据训练的机器学习模型的解释可以揭示潜在的反应机制。在本文中，我们讨论了自动驾驶实验室的不同元素，并介绍了最近在自主反应建模和优化方面的努力。进一步发展和采用机器学习引导的闭环实验策略可以充分发挥自主化学科学和工程的潜力，加速先进材料和分子的发现和发展。
Recent advances in machine learning (ML) and artificial intelligence have provided an exciting opportunity to computerize the fundamental and applied studies of complex reaction systems via self-driving laboratories. Autonomous robotic experimentation can enable time-, material-, and resource-efficient exploration and/or optimization of high-dimensional space reaction systems. Furthermore, interpretation of the ML models trained on the experimental data can unveil the underlying reaction mechanisms. In this article, we discuss different elements of a self-driving lab, and present recent efforts in autonomous reaction modeling and optimization. Further development and adoption of ML-guided closed-loop experimentation strategies can realize the full potential of autonomous chemical science and engineering to accelerate the discovery and development of advanced materials and molecules.