Course Outline
Introduction
- Learning through Positive Reinforcement for Government Applications
Elements of Reinforcement Learning for Government
Important Terms (Actions, States, Rewards, Policy, Value, Q-Value, etc.) for Government Operations
Overview of Tabular Solution Methods for Government Use
Creating a Software Agent for Government Applications
Understanding Value-based, Policy-based, and Model-based Approaches in Government Contexts
Working with the Markov Decision Process (MDP) for Government Operations
How Policies Define an Agent's Way of Behaving for Government Services
Using Monte Carlo Methods for Government Analysis
Temporal-Difference Learning for Government Applications
n-step Bootstrapping in Government Contexts
Approximate Solution Methods for Government Use
On-policy Prediction with Approximation for Government Operations
On-policy Control with Approximation for Government Services
Off-policy Methods with Approximation for Government Applications
Understanding Eligibility Traces in Government Contexts
Using Policy Gradient Methods for Government Operations
Summary and Conclusion for Government Use
Requirements
- Experience with machine learning for government applications
- Programming experience
Audience
- Data scientists working in or with government agencies
