Frequent action advice from a simulated teacher speeds up deep reinforcement learning in a visually aliased Minecraft maze, with persistent advice working best.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2019 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Improving Deep Reinforcement Learning in Minecraft with Action Advice
Frequent action advice from a simulated teacher speeds up deep reinforcement learning in a visually aliased Minecraft maze, with persistent advice working best.