Deep Reinforcement Learning for Industrial Insertion Tasks with Visual Inputs and Natural Rewards.
Gerrit SchoettlerAshvin NairJianlan LuoShikhar BahlJuan Aparicio OjeaEugen SolowjowSergey LevinePublished in: IROS (2020)
Keyphrases
- reinforcement learning
- visual tasks
- markov decision processes
- transfer learning
- sensory inputs
- function approximation
- visual information
- real world
- learning algorithm
- model free
- state space
- visual features
- temporal difference
- reinforcement learning algorithms
- optimal policy
- supervised learning
- industrial applications
- visual perception
- dynamic programming
- previously learned
- input data
- optimal control
- real robot
- control policy
- multi agent
- multi agent environments