Reinforcement Learning and Minecraft - ML Con Spring 2018
MLCon | Machine Learning Conference via YouTube
Overview
Explore reinforcement learning and its application in Minecraft through the Malmo project in this 41-minute conference talk from ML Con 2018 Spring. Discover how Lars Gregori from SAP Customer Experience utilizes Microsoft's AI test environment to solve complex problems using (Deep) Reinforcement Learning techniques. Learn about Minecraft's potential as an experimental AI platform, understand the basics of supervised and unsupervised learning, and delve into practical examples such as the "Falling to Lava" scenario. Gain insights into deep reinforcement learning, image subtraction, and classification methods, complete with code demonstrations and valuable takeaways for implementing these concepts in your own projects.
Syllabus
Intro
Minecraft
What is Minecraft
Creative and building aspects
Project Mountain
Supervised
Unsupervised
Agent
Trial and Error
Reinforcement Learning
Falling to Lava
Matrix
Formula
Example
Deep Reinforcement Learning
Subtracting Images
Classification
Code
Demo
Takeaway
Book
GitHub
Taught by
MLCon | Machine Learning Conference