Membrane Science Meets Machine Learning: Critical Factors and Practices on ML Assisted Membrane Fabrication

The evolving membrane technology integrated with machine learning (ML) algorithm can significantly advance the novel membrane material design and fabrication. Membrane fabrication is a multivariate operation that involves selection and optimization of various parameters. Polymers have been a potential proven material for membrane fabrication. Recently, the large data in polymer membranes in literature is explored and interpreted by machine learning (ML) models for novel membrane fabrications. With these ML models, an efficient method is devised other than the age-old trials and errors for finding desirable separation performance. A two-way information gateway is necessary to achieve the desired objective of designing high performance membrane. Whereby experienced researchers and data scientists from both sides need to provide valuable insights into novel membrane development process. In this study, we offer a midway platform by analyzing reported ML-assisted membrane fabrications via lensing through the overall ML development. This work culminates in identifying four crucial factors affecting ML-assisted membrane development: data mining, material functional description, selection of ML models, and model interpretation. Furthermore, three novel research practices with ML assisted membrane material selection, modification and filler production were discussed in detail. We believe the proposed approaches and analysis through our identified critical factors with our three explored work will prove crucial for the future of ML-assisted membrane material design, fabrication, and development.