An Investigation into Pre-Training Object-Centric Representations for Reinforcement Learning.
Jaesik YoonYi-Fu WuHeechul BaeSungjin AhnPublished in: CoRR (2023)
Keyphrases
- reinforcement learning
- supervised learning
- object description
- function approximation
- object model
- training set
- multiple objects
- d objects
- object representation
- training samples
- training process
- complex objects
- dynamic programming
- function approximators
- test set
- markov decision processes
- target object
- training algorithm
- object models
- representation scheme
- reinforcement learning algorithms
- object segmentation
- spatial relations
- data objects
- neural network
- keypoints
- state space
- hidden markov models
- learning process
- multi agent
- e learning
- learning algorithm
- machine learning