HyperDQN: A Randomized Exploration Method for Deep Reinforcement Learning.
Ziniu LiYingru LiYushun ZhangTong ZhangZhi-Quan LuoPublished in: ICLR (2022)
Keyphrases
- high accuracy
- reinforcement learning
- computational complexity
- dynamic programming
- detection method
- theoretical analysis
- support vector machine
- preprocessing
- objective function
- high precision
- synthetic data
- experimental evaluation
- computational cost
- neural network
- detection algorithm
- optimization algorithm
- clustering method
- state space
- function approximation
- segmentation algorithm
- support vector machine svm
- color images
- multiresolution
- learning process
- multiscale
- learning algorithm