Artificial Intelligence in Science and Engineering

Preface

1 Artificial Intelligence and Complex Systems: What It Can and Cannot Do
1.0 Introduction
1.1 A Glance at History
1.2 Complex Media and Systems
1.3 Three Types of Complex Systems
1.4 Physics-Informed and Data-Driven Approach to Complex Media and Phenomena
1.5 What Artificial Intelligence Cannot Do

2 Neural Networks and Other Machine-Learning Algorithms
2.0 Introduction
2.1 Training of Neural Networks: Backpropagation
2.2 Classification of Learning
2.3 Weak Learners and Boosting Algorithms
2.4 Activation Functions
2.5 Types of Neural Networks

2.6 Training of Large Neural Networks
2.7 Other Machine-Learning Algorithms
2.8 Methods of Minimizing the Loss Function
2.9 Challenges and Future Directions

3 Solving Differential and Partial Differential Equations
3.0 Introduction
3.1 Solving Ordinary Differential Equations
3.2 Solving Partial Differential Equations
3.3 Solving High-Dimensional Partial Differential Equations: Deep BSDE Algorithm
3.4 Feynman-Kac Solution for Backward Kolmogorov Equation of Stochastic Processes
3.5 Data-Driven Discretization of Partial Differential Equations
3.6 Other Models
3.7 Space-Time Fractional Partial Differential Equations
3.8 Challenges and Future Directions

4 Fluid Mechanics: Single-Phase Flow
4.0 Introduction
4.1 The Microscopic Conservation Laws
4.2 A Glance at History
4.3 Kinematics of Fluid Flow
4.4 Dynamics of Fluid Flow
4.5 Modeling Flow Systems of Type I
4.6 Data-Driven Neural Networks for Flow Systems of Type I
4.7 Physics-Informed and Data-Driven Machine-Learning Approach
4.8 Turbulent Flows
4.9 Control of a Flow Field
4.10 Aerodynamic Systems
4.11 Machine Learning for Accelerating Direct Numerical Simulations
4.12 Challenges and Future Directions

5 Fluid Mechanics: Multiphase Flows
5.0 Introduction
5.1 Physics-Informed Simulation of Two-Phase Flows
5.2 Data-Driven Approach to Simulating Two-Phase Flows
5.3 Multiphase Flow in Heterogeneous Porous Materials and Media
5.4 Challenges and Future Directions

6 Heat and Mass Transfer Processes
6.0 Introduction
6.1 Heat and Mass Transfer Processes
6.2 Applications of Neural Networks to Heat Transfer Processes
6.3 Mass Transfer
6.4 Challenges and Future Directions

7 Porous Materials and Media
7.0 Introduction
7.1 Characterization of Core-Scale Porous media
7.2 Characterization of Large-Scale Porous Media
7.3 Reconstruction of Porous Media
7.4 Data-Driven Neural Networks for Simulating Single-Phase Flow and Transport Processes
7.5 Physics-Informed Neural Networks for Simulating Single-Phase Flow and Transport
7.6 Two-Phase Flow
7.7 Thermo-Hydro-Mechanical Processes
7.8 Data-Driven Neural Networks for Two-Phase Flow
7.9 Challenges and Future Directions

8 Porous Materials and Media
8.0 Introduction
8.1 Quantum Monte Carlo Method
8.2 First-Principle Simulation: Density-Functional Theory Calculations
8.3 Molecular Dynamics Simulation
8.4 Active Learning
8.5 Other Aspects of Development of Force Fields by Machine Learning Algorithms
8.6 Challenges and Future Directions

9 Membranes for Separation of Fluid Mixtures
9.0 Introduction
9.1 Data-Driven Neural Networks for Separation Processes
9.2 Data-Driven Approach for Designing and Screening of Membranes? Materials
9.3 Application of Generative Adversarial Networks to Membrane Separation
9.4 Data-Driven Neural Network for Minimizing Membrane Fouling
9.5 Physics-Informed Modeling of Flow in Membranes
9.6 Challenges and Future Directions

10 Catalysis and Reaction Engineering
10.0 Introduction
10.1 Data-Driven Machine-Learning Algorithms for Predicting Catalytic Activity and Yield
10.2 Data-Driven Machine-Learning Algorithms for Design and Optimization of New Catalysts
10.3 Data-Driven Neural Networks for Predicting Potential Energy Surface in Catalysis
10.4 Applications of Behler-Parrinello Generalized Neural-Network Representation of High-Dimensional Potential Energy Surfaces
10.5 Machine-Learning Approach for Discovering and Designing New Catalysts Using Density-Functional Theory Data
10.6 Machine-Learning Algorithms for Identifying Catalytic Reaction Networks
10.7 Black-Box, Grey-Box, and Glass-Box Methods
10.8 Challenges and Future Directions

11 Materials Science
11.0 Introduction
11.1 Machine-Learning Approach for Designing Polymers and other Macromolecules
11.2 Machine-Learning Algorithms for Crystalline Solids
11.3 Generative Approach for Inverse Modeling of Material Discovery
11.4 Materials Interface
11.5 Challenges and Future Directions

12 Materials Science
12.0 Introduction
12.1 Molecular Dynamics Simulation
12.2 Machine Learning for Coarse-Grained Force Fields
12.3 Machine-Learning Evolutionary Approach to Predicting Protein Structure
12.4 Neural Network Approach to Protein Structure
12.5 Challenges and Future Directions

13 Drug Discovery
13.0 Introduction
13.1 Machine-Learning Methods for Quantitative Structure-Property Relationships
13.2 Machine-Learning Approach for Design of Antibacterial and Antimicrobial Peptides
13.3 Recurrent Neural Network Model for Drug Design
13.4 Design of Proteins for Neutralizing Lethal Snake Venom
13.5 Drugs for Viral Proteins of SARS-Cov-2
13.6 Machine Learning for Predicting Drug-Target Interactions
13.7 Challenges and Future Directions

14 Medical Imaging and Anatomical Diagnosis
14.0 Introduction
14.1 Image Classification
14.2 Detection
14.3 Segmentation
14.4 Registration
14.5 Image Enhancement
14.6 Extracting Features from a Medical Image
14.7 Leveraging Textual Reports to Improve Classification of Medical Images
14.8 Textual Description of Medical Images
14.9 Anatomical Applications
14.10 14.10 Commonalities of Images of Porous Media and Biological Organs
14.11 Challenges and Future Directions

15 Environmental and Climate Sciences
15.0 Introduction
15.1 Hydrology

15.2 Transport of Contaminants in Groundwater
15.3 Soil Moisture
15.4 Carbon Dioxide Storage in Porous Formations
15.5 Climate Models
15.6 Challenges and Future Directions

16 Learning Governing Equations for Datasets
16.0 Introduction
16.1 Type-II Systems
16.2 Type-III Systems
16.3 Challenges and Future Directions