P
Initializing...
Finite-sum form of the bias bound for the reward-corrected estimator, valid for every $\gamma\ge0$ · Prove2Me