Worst-case Performance of Greedy Policies in Bandits with Imperfect Context Observations

Open in new window