We analyze optimal monetary policy when a central bank has to learn about an unknown coefficient that determines the effect of surprise inflation on aggregate demand. We derive the optimal policy under active learning and compare it to two limiting cases---certainty equivalence policy and...