Quasi-Newton Optimization in Deep Q-Learning for Playing ATARI Games.
Jacob RafatiRoummel F. MarciaPublished in: CoRR (2018)
Keyphrases
- quasi newton
- optimization method
- optimization methods
- game playing
- step size
- newton method
- reinforcement learning
- optimization procedure
- state space
- imperfect information
- multi agent
- function approximation
- optimization algorithm
- cooperative
- game theory
- neural network
- learning rate
- optimization problems
- dynamic programming
- machine learning
- neural network training
- model free
- convergence speed