Optimistic Thompson Sampling reduces regret in unknown multi-player games.
problem Navigating uncertainty in unknown multi-player games with strategic decision-making.
method Introduces Thompson Sampling algorithms that exploit opponents' actions and reward structures.
result Achieves over tenfold improvements in experimental budgets with logarithmic regret bound.