Proposal of Ephemeral Value Adjustment with Dimensionality Reduction in Deep Reinforcement Learning.
Daiki KuyoshiYuta SuzukiSatoshi YamanePublished in: GCCE (2021)
Keyphrases
- dimensionality reduction
- reinforcement learning
- high dimensional data
- high dimensional
- function approximation
- principal component analysis
- data points
- low dimensional
- pattern recognition
- pattern recognition and machine learning
- feature extraction
- manifold learning
- high dimensionality
- feature selection
- temporal difference
- feature space
- optimal policy
- reinforcement learning algorithms
- linear discriminant analysis
- machine learning
- input space
- markov decision processes
- dimensionality reduction methods
- multi agent
- data representation
- locally linear embedding
- learning algorithm
- state space
- metric learning
- robotic control
- structure preserving
- temporal difference learning
- lower dimensional
- dynamical systems
- model free
- learning problems
- transfer learning
- kernel pca
- text categorization
- deep learning
- action space
- distance measure
- multi agent reinforcement learning
- random projections