In a two-player bandit game where players see each other's actions but not rewards, competition reduces exploration, cooperation increases it, and neutral players can outperform a single player by observing each other.
Toward a theory of discounted repeated games with imperfect monitoring
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.GT 1years
2019 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Multiplayer Bandit Learning, from Competition to Cooperation
In a two-player bandit game where players see each other's actions but not rewards, competition reduces exploration, cooperation increases it, and neutral players can outperform a single player by observing each other.