Found 1 document(s)
Page 1 of 1
-
We propose the first regret-based approach to the Graphical Bilinear Bandits problem, where n agents in a graph play a stochastic bilinear bandit game with each of their neighbors. This setting reveals a combinatorial NP...