Learning Sequential Decision Tasks for Robot Manipulation with Abstract Markov Decision Processes and Demonstration-Guided Exploration.
David KentSiddhartha BanerjeeSonia ChernovaPublished in: Humanoids (2018)
Keyphrases
- markov decision processes
- sequential decision tasks
- reinforcement learning
- model based reinforcement learning
- state space
- complex environments
- optimal policy
- guided exploration
- learning algorithm
- dynamic programming
- learning process
- finite state
- learning tasks
- mobile robot
- real time dynamic programming
- machine learning
- computer simulation
- markov chain
- reinforcement learning algorithms
- partially observable
- policy iteration
- supervised learning
- transition matrices