Machine Learning in Games: Training AI Agents to Race Using Reinforcement Learning @NVIDIAGameDeveloper
Machine Learning in Games: Training AI Agents to Race Using Reinforcement Learning @NVIDIAGameDeveloper
Uploaded January 2026 | Updated September 2026, 2 weeks ago
In this session replay, we explore the process of teaching an AI agent to drive a vehicle as fast as possible around a racetrack in a cutting-edge video game, using reinforcement learning techniques built on TensorFlow with NVIDIA GPU support. From exploring some early prototypes and experimenting with different features to integrating the system into a full production codebase and optimizing the training, we'll describe how we achieved our goal: an AI driver as fast as the best players.
Machine Learning in Games: Training AI Agents to Race Using Reinforcement LearningNeural Shading Course: Part 2 - How Neural Shading Solves Common Rendering ProblemsNsight Graphics 2023.4 - Release SpotlightBonsai Diorama Demo | NVIDIA RTX Unreal Engine 5.6.1Scaling Game Performance Analysis with NVIDIA NsightBuilding AI-Powered NPCs in Unreal Engine 5.7Neural Shading Course: Part 9 - Get Your Game Production Ready with Neural ShadingQUAKE 2 on Vulkan BackstoryNVIDIA ACE | OurPalm Partner SpotlightIntroducing NVIDIA RTX Kit: Transforming Rendering with AI and Path TracingNsight Graphics 2018.6{Teaser} Getting Started with Nsight Developer Tools | GTC Fall 2022
NVIDIA Game Developer |

Machine Learning in Games: Training AI Agents to Race Using Reinforcement Learning

SHARE TO X SHARE TO REDDIT SHARE TO FACEBOOK WALLPAPER