| About this Abstract |
| Meeting |
MS&T24: Materials Science & Technology
|
| Symposium
|
Frontiers of Machine Learning on Materials Discovery
|
| Presentation Title |
Machine-Learning-Aided Discovery of Metal-Organic Frameworks for Water Harvesting |
| Author(s) |
Li-Chiang Lin, Zhi-Xun Xu, Shiue-Min Shih, Yi-Ming Wang, I-Ting Sung |
| On-Site Speaker (Planned) |
Li-Chiang Lin |
| Abstract Scope |
Harvesting atmospheric water via adsorption using metal-organic frameworks (MOFs) has drawn considerable attention. A key to the success of such technology is the selection of optimal adsorbents. To date, >10,000 MOFs have been reported experimentally, while orders of magnitude more candidates have been theoretically predicted. Given the large materials space, employing machine-learning-aided discovery approaches can play a critical role. Specifically, incorporated with large-scale molecular simulations, machine learning models are developed. In this study, widely adopted tree-based methods as well as convolution neural networks (CNNs) are employed. For the former, aside from including commonly used features such as largest cavity diameter and surface area, newly developed descriptors are proposed for enhanced prediction performance. For the latter, computer vision techniques are exploited to “see” the structures for training and prediction. We anticipate the outcomes of this work can facilitate future efforts on the development of optimal water adsorbents. |