DiffGen: Robot Demonstration Generation via Differentiable Physics Simulation, Differentiable Rendering, and Vision-Language Model

Generating robot demonstrations through simulation is widely recognized as an effective way to scale up robot data. Previous work often trained reinforcement learning agents to generate expert policies, but this approach lacks sample efficiency. Recently, a line of work has attempted to generate robot demonstrations via differentiable simulation, which is promising but heavily relies on reward design, a labor-intensive process. In this paper, we propose DiffGen, a novel framework that integrates differentiable physics simulation, differentiable rendering, and a vision-language model to enable automatic and efficient generation of robot demonstrations. Given a simulated robot manipulation scenario and a natural language instruction, DiffGen can generate realistic robot demonstrations by minimizing the distance between the embedding of the language instruction and the embedding of the simulated observation after manipulation in representation space. The embeddings are obtained from the vision-language model, and the optimization is achieved by calculating and descending gradients through the differentiable simulation, differentiable rendering, and vision-language model components. Experiments demonstrate that with DiffGen, we could efficiently and effectively generate robot data with minimal human effort or training time. The videos of the results can be accessed at https://sites.google.com/view/diffgen.

Authors

• Yang Jin
• Jun Lv
• Shuqiang Jiang
• Cewu Lu

Keywords

• Training
• Visualization
• Natural languages
• Reinforcement learning
• Rendering (computer graphics)
• Robots
• Physics
• Optimization
• Intelligent robots
• Videos
• Physical Simulation
• Differentiable Rendering
• Vision-language Models
• Natural Language
• Language Teaching
• Efficient Generation
• Line Of Work
• Robot Manipulator
• Automatic Generation
• Objective Function
• Gradient Descent
• Crucial Step
• Sequence Of Actions
• System Identification
• Sidewall
• Visual Observation
• Representation Learning
• Language Model
• Optimization Step
• Task Instructions
• Reward Function
• Simulation Integration
• State St
• Learning Rate Of 1e
• Reward Learning
• Physics-based Simulation
• Goal State
• Robotic Tasks
• Training Policy