# Physics Simulation

Genesis is a comprehensive physics simulation platform designed for robotics, embodied AI, and physical AI applications. It is a general-purpose physics engine built from the ground up to simulate a wide range of materials and physical phenomena. As a lightweight, ultra-fast, Pythonic, and user-friendly simulation platform, it also features a powerful realistic rendering system and a data generation engine that converts natural language descriptions into various data modalities. By integrating its core physics engine, Genesis enhances the upper-level generative agent framework, aimed at achieving fully automated data generation for robotics and beyond.

Oasis is the first playable, real-time, open-world AI model developed by Decart AI. It is an interactive video game entirely generated end-to-end using a Transformer, based on frame-by-frame generation. Oasis can receive user keyboard and mouse inputs to generate gameplay in real-time, internally simulating physics, game rules, and graphics. The model learns from direct gameplay observations, enabling users to move, jump, pick up items, destroy blocks, and more. Oasis is viewed as the first step toward researching more complex interactive worlds and may potentially replace traditional game engines in the future. Its implementation requires improvements in model architecture and breakthroughs in inference technology to facilitate real-time interaction with users. Decart AI employs the latest diffusion training and Transformer modeling methods, combining them with large language models (LLMs) to train an autoregressive model capable of generating video based on user actions. In addition, Decart AI has developed a proprietary inference framework to maximize the utilization of the NVIDIA H100 Tensor Core GPU and support the upcoming Sohu chip from Etched.

ProtoMotions is a project dedicated to creating interactive physics-based virtual agents. It supports IsaacGym and IsaacSim and is built on Hydra and OmegaConfig, facilitating simple configuration management. This project offers researchers and developers a platform to develop and test physics-based character animation techniques, applicable not only to academic research but also in fields such as gaming, film, and virtual reality.

Video2Game is a technology that can transform a single video into a high-quality virtual environment with real-time interactivity, realism, and browser compatibility. It achieves high-quality surface geometry by constructing a large-scale NeRF model, which is then converted into a grid representation with corresponding rigid body dynamics to support interaction. Using UV-mapped neural textures ensures both richness of expression and compatibility with game engines. The final result is a virtual environment where virtual characters can interact, respond to user control, and provide high-resolution rendering from new camera perspectives in real time.