Learning Actionable Representations from Visual Observations

In this work we explore a new approach for robots to teach themselves about the world simply by observing it. In particular we investigate the effectiveness of learning task-agnostic representations for continuous control tasks. We extend Time-Contrastive Networks (TCN) that learn from visual observations by embedding multiple frames jointly in the embedding space as opposed to a single frame. We show that by doing so, we are now able to encode both position and velocity attributes significantly more accurately. We test the usefulness of this self-supervised approach in a reinforcement learning setting. We show that the representations learned by agents observing themselves take random actions, or other agents perform tasks successfully, can enable the learning of continuous control policies using algorithms like Proximal Policy Optimization (PPO) using only the learned embeddings as input. We also demonstrate significant improvements on the real-world Pouring dataset with a relative error reduction of 39. 4% for motion attributes and 11. 1% for static attributes compared to the single-frame baseline. Video results are available at https://sites.google.com/view/actionablerepresentations.

Authors

• Debidatta Dwibedi
• Jonathan Tompson
• Corey Lynch
• Pierre Sermanet

Keywords

• Task analysis
• Robots
• Visualization
• Reinforcement learning
• Aerospace electronics
• Solid modeling
• Semantics
• Visual Observation
• Representation Learning
• Latent Space
• Control Task
• Single Frame
• Embedding Learning
• Random Action
• Proximal Policy Optimization
• Time Step
• Visual Representation
• Real-world Data
• Latent Factors
• Video Frames
• True State
• State Representation
• Object Position
• Robot Control
• Self-supervised Learning
• Policy Learning
• Robust Representation
• Object Velocity
• Imitation Learning
• Inverse Reinforcement Learning
• Robotic Tasks
• Cheetah
• Supervisory Signal
• Convolutional Feature Maps
• Alignment Errors
• Simulation Environment
• Convolutional Neural Network