Topological Dynamics in Metamodel Discovery with Artificial Intelligence: From Biomedical to Cosmological Technologies

Cover
Half Title
Series Page
Title Page
Copyright Page
Dedication
Table of Contents
Preface
About the Author
Part I: Fundamentals
Chapter 1: Artificial Intelligence and Dynamical Systems
1.1 Artificial Intelligence for Model Discovery
1.2 Primer on Deep Learning
1.3 Neural Networks as Models for Dynamical Systems
1.4 Deep Learning with Biomedical Applications
1.5 Convolutional Neural Networks in Drug Design
1.6 Deep Learning for Dynamic Targets in Biomedicine
1.7 Deep Learning Has Solved One of the Protein Folding Problems
1.8 AI-Empowered Metamodel Discovery for Hierarchical Dynamical Systems: Adiabatic Regimes, Latent Manifolds, Quotient Spaces
1.9 Metadynamics for Metamodels: Mapping Out the Quotient Manifold with a Dedicated Autoencoder
1.10 Metamodels for the Digital Mind
References
Chapter 2: Topological Methods for Metamodel Discovery with Artificial Intelligence
2.1 AI-Based Metamodel Discovery for Hierarchically Complex Dynamical Systems
2.2 Autoencoders of Latent Coordinates in Dynamical Systems
2.3 Deep Learning Scheme to Discover Underlying Differential Equations from Time Series
2.4 Autoencoders for Molecular Dynamics of Biological Matter
2.5 Topological Dynamics on Latent Manifolds: Metamodels without Equations
2.6 Unraveling Topological Quotient Spaces for Dynamical Systems Metamodeling
2.7 Learning to Encode and Propagate Topological Dynamics: Metamodel for Ubiquitin Folding in the Cell
2.8 Metamodels for Hierarchical Dynamics Discovered through Autoencoder Batteries
References
Part II: Applications
Chapter 3: Artificial Intelligence Reverse-Engineers In Vivo Protein Folding
3.1 Deconstructing In Vivo Protein Folding: Topological Metamodel Created by Transformer Technology
3.2 Empowering Molecular Dynamics with Transformer Technology
3.2.1 Propagating the Topological Dynamics in Textual Form with a Transformer Neural Network
3.2.2 Topological Metamodeling Requires Two Autoencoders and a Transformer
3.3 Protein Folding as a Textually Encodable Dynamical Metamodel: Mathematical Validation
3.4 Injecting In Vivo Reality into the Transformer-Generated Metamodel
3.5 Propagating In Vitro Folding Pathways
3.6 Atomistic MD Simulation of an In Vivo Folding Setting
3.7 Propagation of In Vivo Folding Trajectories Using Transformer Technology
3.8 The AI Platform to Generate In Vivo Folding Pathways
3.9 Reverse-Engineering the Expeditious In Vivo Context I: Iterative Annealing in the Apo GroEL Chamber
3.10 Reverse-Engineering the In Vivo Context II: GroEL Chamber in the (ATP) 7 State
3.11 In Vivo Pathways for Protein Folding: AI Metamodels Live Up to the Challenge
3.12 Metamodels with Implicit Content: Co-translational Protein Folding
References
Chapter 4: The Drug-Induced Protein Folding Problem: Metamodels for Dynamic Targeting
4.1 Protein Structure is a Dynamic Object: Lesson for Targeted Therapy
4.2 Deep Learning to Target Moving Targets in Molecular Therapy
4.3 AI-Based Metamodel to Infer Drug-Induced Folds in Targeted Proteins
4.4 Learning to Induce Folds in Targeted Proteins
4.5 Experimental Corroboration of the Dynamic Metamodel for Drug-Induced Folding
4.6 A Topological Metamodel of the Induced Folding Dynamics Corroborates the Experimentally Validated Structural Adaptation in the Drug/Target WBZ_4/JNK Complex
4.7 AI Teaches Drug Designers How to Target Proteins by Exploiting a Dynamic Metamodel of Induced Folding
References
Chapter 5: Targeting Protein Structure in the Absence of Structure: Metamodels for Biomedical Applications
5.1 Therapeutic Disruption of Dysfunctional Protein Complexes in the Absence of Reported Structure
5.2 AI Guides the Therapeutic Disruption of a Dysfunctional Complex with No Reported Structure
5.3 Regulatory Sites in the Epistructure of a Protein Target
5.4 AI-Based Metamodel to Infer Regulation-Modulated Epitopes in the Absence of Target Structure
5.5 Structural Metamodel for Therapeutic Disruption of Dysfunctional Deregulated Protein Complexes in the Absence of Structure
5.6 Experimentally Validating the Dynamic Metamodel of the Target Protein Structure by Developing a Molecular Targeted Therapy for Heart Failure
5.7 The Dynamic Metamodel of Protein Structure Enables Discovery OF Biological Cooperativity
References
Chapter 6: Autoencoder as Quantum Metamodel of Gravity: Toward an AI-Based Cosmological Technology
6.1 The Quest for Quantum Gravity
6.2 Quantum Gravity Autoencoder for a Neural Network with Emergent Gravity
6.3 Relativistic Strings-Turned Quanta in Machine Learning Physics
6.4 The Universe as a Variational Autoencoder
6.5 Quantum Gravity Autoencoders and the Origin of the Universe
6.6 Technologies for Cosmological Manipulation Leveraging Quantum Gravity Autoencoders
References
Epilogue
E.1 Topological Metamodels Breed Computational Intuition
E.2 AI Probes an Equivalence between “Wormhole” and Quantum Entanglement
E.3 Emergent Space-Time Topology Created by Entangling Quantum-Gravity Autoencoders
E.4 Metamodel Discovery with Extra Dimensions and Lost Symmetries: Artificial Intelligence Deconstructs Quantum Mechanics
References
Appendix
A.1 Code for Dehydron Identification
A.2 Machine Learning Method to Infer Structure Wrapping and Dhydron Pattern in the Absence of Protein Structure: The Twilighter
A.3 AI Platform to empower Molecular Dynamics
A.3.1 Dynamical Feature Extraction for AI-Empowered Protein Folding Simulations
A.3.2 How to Extend Molecular Dynamics Simulations
A.3.3 AI-Generated Protein Folding Pathways
A.3.4 Molecular Dynamics Run from an AI Platform
A.4 Protein Folding Pathway Generated by Transformer-Engendered Topological Dynamics
A.5 Endowing the Quantum Mechanics Autoencoder with Emergent Gravity
A.6 Incorporationg an Extra Dimension in Space-Time through an Autoencoder of the Standard Model: Decoding the Higgs Mechanism
References
Index