Can machine learning methods guide gas separation membranes fabrication?

Tayyebi, Arash; Alshami, Ali; Yu, Xue; Kolodka, Edward

doi:10.1016/j.memlet.2022.100033

Cited by 19 publications

(9 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Compared to molecular descriptors, fingerprinting methods provide a more dynamic representation that encompasses the characteristics of materials through their fragment features [ 25 ]. There are various types of molecular fingerprints, which are determined by the method used to convert the molecular fragment into a binary string [ 26 ]. Fingerprints with longer bit strings are more reliable for a similarity search since each significant bond in a molecule is defined separately as a sequence of binary digits (bits), and they have more stored information regarding the molecular properties.…”

Section: Methodsmentioning

confidence: 99%

Prediction of organic compound aqueous solubility using machine learning: a comparison study of descriptor-based and fingerprints-based models

Tayyebi,

Alshami,

Rabiei

et al. 2023

J Cheminform

Self Cite

View full text Add to dashboard Cite

A reliable and practical determination of a chemical species’ solubility in water continues to be examined using empirical observations and exhaustive experimental studies alone. Predictions of chemical solubility in water using data-driven algorithms can allow us to create a rationally designed, efficient, and cost-effective tool for next-generation materials and chemical formulations. We present results from two machine learning (ML) modeling studies to adequately predict various species’ solubility using data for over 8400 compounds. Molecular-descriptors, the most used method in previous studies, and Morgan fingerprint, a circular-based hash of the molecules' structures, were applied to produce water solubility estimates. We trained all models on 80% of the total datasets using the Random Forest (RFs) technique as the regressor and tested the prediction performance using the remaining 20%, resulting in coefficient of determination (R2) test values of 0.88 and 0.81 and root-mean-square deviation (RMSE) test values 0.64 and 0.80 for the descriptors and circular fingerprint methods, respectively. We interpreted the produced ML models and reported the most effective features for aqueous solubility measures using the Shapley Additive exPlanations (SHAP) and thermodynamic analysis. Low error, ability to investigate the molecular-level interactions, and compatibility with thermodynamic quantities made the fingerprint method a distinct model compared to other available computational tools. However, it is worth emphasizing that physicochemical descriptor model outperformed the fingerprint model in achieving better predictive accuracy for the given test set.

show abstract

Section: Methodsmentioning

confidence: 99%

Prediction of organic compound aqueous solubility using machine learning: a comparison study of descriptor-based and fingerprints-based models

Tayyebi,

Alshami,

Rabiei

et al. 2023

J Cheminform

Self Cite

View full text Add to dashboard Cite

show abstract

“…A ngerprinting method is a more dynamic representation than molecular descriptors and can cover materials by their fragments features. Molecular ngerprints have several types based on the method by which the molecular fragment is transformed into a bit string [24]. Fingerprints with longer bit strings are more reliable for a similarity search since each signi cant bond in a molecule is de ned separately as a sequence of binary digits (bits), and they have more stored information regarding the molecular properties.…”

Section: Data Preparationmentioning

confidence: 99%

Prediction of Organic Compound Aqueous Solubility Using Interpretable Machine Learning- A Comparison Study of Descriptor-Based and Topological Models

Tayyebi

Alshami

Rabiei

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

A reliable and practical determination of a chemical species’ solubility in water continues to be examined using empirical observations and exhaustive experimental studies alone. Predictions of chemical solubility in water using data-driven algorithms can allow us to create a rationally designed, efficient, and cost-effective tool for next-generation materials and chemical formulations. We present results from two machine learning (ML) modeling studies to adequately predict various species’ solubility using data for over 8,400 compounds. Molecular-descriptors, the most used method in previous studies, and Morgan fingerprint, a topological, circular-based hash of the molecules' structures, were applied to produce water solubility estimates. We trained all models on 80% of the total datasets using the Random Forest (RFs) technique as the regressor and tested the prediction performance using the remaining 20%, resulting in R2 test values of 0.88 and 0.82 for the descriptors and circular fingerprint methods, respectively. We interpreted the produced ML models and reported the most effective features for aqueous solubility measures using Shapley Additive exPlanations (SHAP) and thermodynamic analysis. Low error, ability to investigate the molecular-level interactions and compatible with thermodynamic quantities made fingerprint a distinct model compared to other available computational tools.

show abstract

“…For example, computational and experimental efforts have shown that 2D materials with nanopores are more energy-efficient in RO desalination compared with polymeric membranes. , Behind the successful application of porous membranes, the discovery and design process of both polymeric and 2D membranes are based mostly on traditional trial-and-error experimental methods, which are time-consuming and expensive. For polymeric membranes, the process of fabrication and synthesis can significantly affect their properties . However, the large search space constituted by the selection of polymers, solvent, and additives renders traditional experimental methods inefficient in discovering new polymeric membranes .…”

mentioning

confidence: 99%

“…6 However, the large search space constituted by the selection of polymers, solvent, and additives renders traditional experimental methods inefficient in discovering new polymeric membranes. 6 One example of such inefficiency is the marginal RO performance improvement of polymeric membranes since 1990. 7 For 2D material membranes, their performances depend not only on the type of materials 8,9 but also on the size and geometry of the artificial nanopores.…”

mentioning

confidence: 99%

“…For polymeric membranes, the process of fabrication and synthesis can significantly affect their properties . However, the large search space constituted by the selection of polymers, solvent, and additives renders traditional experimental methods inefficient in discovering new polymeric membranes . One example of such inefficiency is the marginal RO performance improvement of polymeric membranes since 1990 .…”

mentioning

confidence: 99%

See 1 more Smart Citation

Machine Learning in Membrane Design: From Property Prediction to AI-Guided Optimization

Cao,

Barati Farimani,

Ock

et al. 2024

Nano Lett.

View full text Add to dashboard Cite

Porous membranes, either polymeric or two-dimensional materials, have been extensively studied because of their outstanding performance in many applications such as water filtration. Recently, inspired by the significant success of machine learning (ML) in many areas of scientific discovery, researchers have started to tackle the problem in the field of membrane design using data-driven ML tools. In this Mini Review, we summarize research efforts on three types of applications of machine learning in membrane design, including (1) membrane property prediction using ML, (2) gaining physical insight and drawing quantitative relationships between membrane properties and performance using explainable artificial intelligence, and (3) ML-guided design, optimization, or virtual screening of membranes. On top of the review of previous research, we discuss the challenges associated with applying ML for membrane design and potential future directions.

show abstract

Can machine learning methods guide gas separation membranes fabrication?

Cited by 19 publications

References 33 publications

Prediction of organic compound aqueous solubility using machine learning: a comparison study of descriptor-based and fingerprints-based models

Prediction of organic compound aqueous solubility using machine learning: a comparison study of descriptor-based and fingerprints-based models

Prediction of Organic Compound Aqueous Solubility Using Interpretable Machine Learning- A Comparison Study of Descriptor-Based and Topological Models

Machine Learning in Membrane Design: From Property Prediction to AI-Guided Optimization

Contact Info

Product

Resources

About