Reactive Exploration to Cope with Non-Stationarity in Lifelong Reinforcement Learning.
Christian Alexander SteinparzThomas SchmiedFabian PaischerMarius-Constantin DinuVihang P. PatilAngela Bitto-NemlingHamid Eghbal-zadehSepp HochreiterPublished in: CoRR (2022)
Keyphrases
- robotic control
- reinforcement learning
- exploration strategy
- exploration exploitation
- non stationary
- reactive agents
- st century
- state space
- action selection
- learning algorithm
- model based reinforcement learning
- competence development
- active exploration
- markov decision processes
- reinforcement learning algorithms
- lifelong learning
- exploration exploitation tradeoff
- neural network
- technology enhanced learning
- optimal control
- function approximation
- learning scenarios
- supervised learning
- dynamic programming
- learning environment
- data mining