Developing AI in Racing Games Using Reinforcement Learning | Unreal Fest Online 2020
Reinforcement learning is about to radically change the development process of AI in video games. In this session, Giuseppe Campana introduces reinforcement learning from a pragmatic point of view. He dives deep into the infrastructure Milestone is developing to produce the AI for its racing games, and show how the studio’s plugin integrates smoothly with Unreal Engine.
Giuseppe explains how Milestone exploits class reflection for analysis and production tools, and how it runs training in a custom minimal game loop.
Great for: game programmers/designers curious about reinforcement learning (no prior knowledge of machine learning is required).
Learn more at http://www.unrealengine.com/games
Other Videos By Unreal Engine
Other Statistics
Unreal Statistics For Unreal Engine
Currently, Unreal Engine has 74,828,456 views for Unreal across 1,300 videos. His channel published over 40 days worth of content for the game, roughly 66.42% of Unreal content that Unreal Engine has uploaded to YouTube.