GameNGen, a neural model-based game engine, is demonstrating the potential to revolutionize how video games are generated and played. An innovative approach developed by Google Research and Tel Aviv University researchers allows for real-time interaction with complex gaming environments without relying on traditional game engines.
As the authors reported, GameNGen can simulate the classic game DOOM at over 20 frames per second, achieving visual quality comparable to the original game.
The core of GameNGen’s functionality lies in its use of diffusion models, a type of generative AI that has become a standard in media generation. The process begins with training a reinforcement learning (RL) agent to play the game, recording its actions and observations. This data is then used to train a diffusion model to predict the next frame based on a sequence of past frames and actions. This method allows the model to simulate complex game state updates, such as managing health and ammo, attacki
We współpracy z: https://cryptoslate.com/ai-model-generates-real-time-playable-doom-with-no-game-engine/
