Message-Dropout: An Efficient Training Method for Multi-Agent Deep Reinforcement Learning.
Woojun KimMyungsik ChoYoungchul SungPublished in: CoRR (2019)
Keyphrases
- reinforcement learning
- multi agent
- high accuracy
- cost function
- synthetic data
- highly efficient
- computational complexity
- training phase
- preprocessing
- significant improvement
- dynamic programming
- computational cost
- feature selection
- segmentation method
- support vector machine svm
- data sets
- probabilistic model
- prior knowledge
- clustering method
- artificial neural networks
- pairwise
- high precision
- objective function
- training process
- feed forward neural networks