A2C deep reinforcement learning reliably fails on four specially designed deceptive games, sometimes learning superstitious behaviors, and its failure modes differ from planning agents.
P.; Brundage, M.; and Bharath, A
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
-
Superstition in the Network: Deep Reinforcement Learning Plays Deceptive Games
A2C deep reinforcement learning reliably fails on four specially designed deceptive games, sometimes learning superstitious behaviors, and its failure modes differ from planning agents.