Materials Informatics I: Methods

Cover
Half Title
Challenges and Advances in Computational: Volume 39
Materials Informatics I: Methods
Copyright
Preface
About This Book
Contents
Editors and Contributors
Part I. Introduction
	1. Introduction to Materials Informatics
		1.1 Introduction
		1.2 ML-Enabled Framework for New Material Design
		1.3 ML-Enabled Framework for Process Optimization
		1.4 Microstructure Informatics
		1.5 Conclusion
		References
	2. Introduction to Cheminformatics for Predictive Modeling
		2.1 Introduction
		2.2 Data in Chemistry
			2.2.1 Data Resources
			2.2.2 Dataset Representation
		2.3 Predictive Modeling Strategies
			2.3.1 Common Tasks
			2.3.2 Algorithms
			2.3.3 Validation
			2.3.4 Interpretation
		2.4 Conclusion and Remarks
		References
	3. Introduction to Machine Learning for Predictive Modeling of Organic Materials
		3.1 Introduction
		3.2 Models Based on Pre-defined Representations
		3.3 Deep Learning (DL)
		3.4 Learning from Scarce Data
			3.4.1 Data Augmentation
			3.4.2 Transfer Learning (TL) and Mutitask Learning (MTL)
			3.4.3 Pre-trained Representations
		3.5 Conclusion
		References
	4. Quantitative Structure–Property Relationships (QSPR) for Materials Science
		4.1 Introduction
		4.2 Different Materials
			4.2.1 Polymers
			4.2.2 Nanomaterials
			4.2.3 Ionic Liquids
			4.2.4 Other Materials
		4.3 Conclusions
		References
Part II. Methods and Tools
	5. Quantitative Structure–Property Relationships (QSPR) and Machine Learning (ML) Models for Materials Science
		5.1 Introduction
		5.2 Theoretical Foundations of QSPR
			5.2.1 Principles of Quantitative Structure–Property Relationships
			5.2.2 Molecular Descriptors and Property Predictions
			5.2.3 Statistical Methods in QSPR Analysis
			5.2.4 Computational Approaches for QSPR Modelling
		5.3 Methodologies in QSPR Studies
			5.3.1 Data Collection and Preprocessing
			5.3.2 Descriptor Selection and Generation
			5.3.3 Model Development and Validation
			5.3.4 Interpretation of QSPR Models
		5.4 Materials Classes
			5.4.1 Nanomaterials
			5.4.2 Catalysts
			5.4.3 Biomaterials
			5.4.4 Polymers (Nonbiological Applications)
		5.5 Recent Advancements and Future Directions
			5.5.1 Machine Learning Approaches in QSPR
			5.5.2 Integration of Multi-scale Modeling
			5.5.3 High-Throughput Screening Techniques
			5.5.4 Emerging Trends in QSPR Research
		5.6 Challenges and Limitations
		5.7 Conclusion
		References
	6. Optimising Materials Properties with Minimal Data: Lessons from Vanadium Catalyst Modelling
		6.1 Introduction
			6.1.1 Material Properties and Small Data
			6.1.2 Small Data Breakthroughs in Chemistry
			6.1.3 Objectives
		6.2 Small Data Optimisation Case Study
			6.2.1 The Role of Vanadium in ESA Catalytic Processes
			6.2.2 Building a Representative Small Dataset
			6.2.3 Visualisation Strategies for Extracting Information from Small Datasets
			6.2.4 Remarks on Minimal Data Strategies
		6.3 Final Remarks
		References
	7. In Silico QSPR Studies Based on CDFT and IT Descriptors
		7.1 Introduction
		7.2 Theoretical Background of CDFT and IT Descriptors
		7.3 Methods
			7.3.1 Computational Details
			7.3.2 Regression Analysis and Machine Learning Approach
		7.4 Applications of CDFT and IT-Based Descriptors Toward QSPR Modeling
			7.4.1 QSPR Studies on Polychlorobiphenyls
			7.4.2 Application of CDFT Descriptors for Toxicity Prediction Against Tetrahymena Pyriformis and Pimephales Promelas
			7.4.3 Application of Information Theory in Chemical Problems
		7.5 Conclusion
		References
	8. Applications of Quantitative Read-Across Structure–Property Relationship (q-RASPR) Modeling in the Field of Materials Science
		8.1 Introduction
		8.2 Application of Informatics in Materials Science
		8.3 A Brief Account on Chemical Similarities
			8.3.1 Fingerprint-Based Similarity
			8.3.2 Distance-Based Similarity
			8.3.3 Physicochemical Property-Based Similarity
			8.3.4 Kernel-Based Similarity
		8.4 The q-RASPR Algorithm and Its Elements
			8.4.1 QSPR
			8.4.2 Read-Across (RA)
			8.4.3 Quantitative Read-Across Structure–Property Relationship (q-RASPR)
		8.5 Sequential Steps for the Development of a q-RASPR Model
		8.6 Applications of q-RASPR
		8.7 Conclusion
		References
	9. Machine Learning Algorithms for Applications in Materials Science I
		9.1 Materials Science and Machine Learning Collaboration: A Foundation for Innovation
			9.1.1 Historical Review: Progress in Machine Learning Algorithms
			9.1.2 A Brief Overview of Technical Concepts of Machine Learning
		9.2 The Transformative Effects of Machine Learning Algorithms in Materials Science
			9.2.1 Different Branches of Materials Science and Presence of ML Algorithms
		9.3 Popular Types of Advanced ML Algorithms for Material Science
			9.3.1 Machine Learning Strategy in Material Sciences: Challenges and Limitations
		9.4 The Success of Machine Learning Strategy in Material Sciences
			9.4.1 Future Directions & New Possibilities in Machine Learning for Materials Science
			9.4.2 Future and Outlook
		References
	10. Machine Learning Algorithms for Applications in Materials Science II
		10.1 Introduction
		10.2 Overview of Machine Learning Methods Used in Material Informatics
			10.2.1 Supervised Learning
			10.2.2 Unsupervised Learning
			10.2.3 Reinforcement Learning
			10.2.4 Transfer Learning
			10.2.5 Deep Learning-Based Models Used in Material Sciences
		10.3 Emerging ML Methods in Material Science
			10.3.1 Explainable AI (XAI) Methods
			10.3.2 Few-Shot Learning (FSL)
		10.4 Case Studies on the Application of Machine Learning in Material Sciences
		10.5 Conclusion
		References
	11. Structure-Property Modeling of Quantum-Theoretic Properties of Benzenoid Hydrocarbons by Means of Connection-Related Graphical Descriptors
		11.1 Introduction
		11.2 Mathematical Preliminaries
			11.2.1 Connection-Based Graphical Invariants
		11.3 Computational Methods
		11.4 Data Analysis
		11.5 Results and Discussion
		11.6 Conclusion
		References
	12. Machine Learning Tools and Web Services for Materials Science Modeling
		12.1 Introduction
		12.2 Machine Learning Tools
			12.2.1 Schrödinger
			12.2.2 Optibrium
			12.2.3 Molecular Operating Environment
			12.2.4 Alvascience
			12.2.5 QSAR-Co-X
			12.2.6 CORAL
			12.2.7 AZOrange
			12.2.8 BioPPSy
			12.2.9 Camb
			12.2.10 eTOXlab
			12.2.11 Open3DQSAR
			12.2.12 ChemML
		12.3 Machine Learning Web Services
			12.3.1 Chemsar
			12.3.2 Online Chemical Modeling Environment
			12.3.3 ChemBench
		12.4 Conclusion
		12.5 Future Directions and Challenges
		References
Index