Deep Reinforcement Learning with Parameterized Action Space for Object Detection.
Zheng WuNaimul Mefraz KhanLei GaoLing GuanPublished in: ISM (2018)
Keyphrases
- action space
- object detection
- reinforcement learning
- state space
- markov decision processes
- state and action spaces
- real valued
- continuous state
- reinforcement learning methods
- action selection
- continuous state spaces
- control policies
- stochastic processes
- computer vision
- function approximators
- function approximation
- reinforcement learning algorithms
- learning algorithm
- optimal policy
- markov decision problems
- state action
- markov decision process
- machine learning
- optimal control
- multi agent
- dynamic programming
- single agent
- control problems
- dynamical systems
- reward function
- heuristic search
- np hard
- search space
- search algorithm
- temporal difference
- policy search
- planning problems