Self-Predictive Dynamics for Generalization of Vision-based Reinforcement Learning.
Kyungsoo KimJeongsoo HaYusung KimPublished in: IJCAI (2022)
Keyphrases
- reinforcement learning
- function approximation
- human computer interaction
- state space
- reinforcement learning algorithms
- dynamic model
- augmented reality
- robotic control
- real time
- vision system
- markov decision processes
- model free
- learning algorithm
- dynamic programming
- optimal policy
- recurrent networks
- stochastic approximation
- social networks
- collective behavior
- function approximators
- sensor fusion
- predictive model
- mobile robot
- multi agent
- dynamical systems
- video sequences
- supervised learning