Deep Learning for Solving and Estimating Dynamic Models in Economics and Finance

An implementation-oriented introduction to deep learning methods for solving and estimating high-dimensional dynamic stochastic models in economics and finance. The exposition is organized around four complementary methodologies: Deep Equilibrium Nets, Physics-Informed Neural Networks, deep surrogate models and Gaussian processes, and GP-based dynamic programming.

Front Matter¶

• Preface

• How to Read This Script

• Notation and Symbols

• Execution Map

Chapters¶

• Introduction to Machine Learning and Deep Learning

• Deep Equilibrium Nets

• The International Real Business Cycle Model

• Neural Architecture Search and Loss Normalization

• Overlapping Generations Models with DEQNs

• Heterogeneous Agents and Young's Method

• Physics-Informed Neural Networks

• Heterogeneous Agent Models in Continuous Time

• Gaussian Processes

• Deep Surrogate Models and Structural Estimation

• Climate Economics and Deep Uncertainty Quantification

• Synthesis and Outlook

Appendices¶

• Glossary

• Matrix Calculus and Automatic Differentiation

• Itô Calculus, Brownian Motion, and Ergodicity

• Fixed Points, Contraction Mappings, and the Bellman Operator

• Reproducibility Information

• Solutions and Guidance for Exercises