Policy Augmentation: An Exploration Strategy For Faster Convergence of Deep Reinforcement Learning Algorithms.
Arash MahyariPublished in: ICASSP (2021)
Keyphrases
- faster convergence
- reinforcement learning algorithms
- exploration strategy
- reinforcement learning
- reinforcement learning problems
- optimal policy
- markov decision problems
- reward function
- markov decision processes
- state space
- markov decision process
- model free
- step size
- convergence speed
- temporal difference
- function approximation
- action selection
- global optimum
- policy iteration
- function approximators
- partially observable
- convergence rate
- unknown environments
- global optimization
- action space
- pso algorithm
- partially observable markov decision processes
- learning algorithm
- particle swarm optimization
- multi agent
- infinite horizon
- transition probabilities
- decision theoretic
- finite state
- linear programming
- supervised learning
- dynamic programming
- machine learning
- decision problems
- learning agent
- mobile robot