Empirical tests in VizDoom show multiple DQN updates per step do not improve performance after learning rate adjustment, with a 4:1 update-to-step ratio optimal before significant degradation.
Deep Reinforcement Learning with Double Q-learning
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2019 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Optimal Use of Experience in First Person Shooter Environments
Empirical tests in VizDoom show multiple DQN updates per step do not improve performance after learning rate adjustment, with a 4:1 update-to-step ratio optimal before significant degradation.