Google and Tel Aviv University Unveil AI Model for Real-Time Doom Simulation
Brief news summary
Researchers from Google and Tel Aviv University have developed GameNGen, an AI model capable of simulating the popular game Doom in real time using AI image generation techniques. This innovative neural network system has the potential to revolutionize real-time video game synthesis by generating graphics on the fly instead of relying on traditional graphical methods. GameNGen can generate new frames of Doom gameplay at a rate of over 20 frames per second using a single tensor processing unit (TPU) optimized for machine learning tasks. Human raters had difficulty distinguishing between actual gameplay footage and GameNGen-generated outputs. The researchers acknowledge that while this development builds upon previous work in the field and aligns with ongoing research in "world models" and "world simulators," there are challenges such as graphical glitches and temporal coherency that need to be addressed. They emphasize that interactive world simulation requires more than just fast video generation.Researchers from Google and Tel Aviv University have introduced GameNGen, an AI model capable of simulating the popular 1993 video game Doom in real time. By leveraging AI image generation techniques from Stable Diffusion, GameNGen serves as a limited game engine, potentially revolutionizing real-time video game synthesis. Instead of using traditional graphical techniques, future games could employ AI engines to generate graphics on the fly. GameNGen can generate new Doom gameplay frames at speeds above 20 frames per second using a single tensor processing unit.
In tests, human raters struggled to differentiate between actual gameplay footage and GameNGen outputs. This concept of "neural rendering" aligns with the ongoing research on AI video synthesis models like OpenAI's Sora. The research paper, titled "Diffusion Models Are Real-Time Game Engines, " explains GameNGen's process by using a modified version of Stable Diffusion 1. 4, an image synthesis diffusion model. However, using Stable Diffusion leads to graphical glitches and challenges with maintaining visual clarity and consistency over time.
Watch video about
Google and Tel Aviv University Unveil AI Model for Real-Time Doom Simulation
Try our premium solution and start getting clients — at no cost to you
