Reinforcement Learning with Videos: Combining Offline Observations with Interaction.
Karl SchmeckpeperOleh RybkinKostas DaniilidisSergey LevineChelsea FinnPublished in: CoRL (2020)
Keyphrases
- video dataset
- reinforcement learning
- action recognition
- video frames
- video sequences
- video data
- real time
- function approximation
- video streams
- human computer interaction
- robotic control
- video analysis
- activity recognition
- state space
- learning algorithm
- neural network
- optimal policy
- user interaction
- human activities
- temporal information
- video content
- multimedia
- video database
- model free
- user generated
- group activities
- sensory inputs