3

We study nonstationary generalized linear bandits (GLBs), where the expected reward is modeled through a nonlinear link function with …