Learning to Navigate in Cities Without a Map

Explore the fascinating world of reinforcement learning applied to urban navigation in this 30-minute lecture from the University of Central Florida. Delve into key concepts such as value functions and actor-critic models, and discover how they can be applied to solve complex courier tasks in unmapped city environments. Examine the problem statement, actions, and environments involved in this challenging scenario. Learn about the architecture and training process behind the solution, including multisite experiments and abolition analysis. Gain insights into the act of creating AI systems for real-world applications, and witness the results through an engaging demo.

Introduction
What is Reinforcement Learning
What is Value Function
Problem Statement
Actions
Environments
Courier task
Architecture
Training
Act of Creating
Actor Critic
Results
Multisite Experiments
Abolition Analysis
Demo