Knowledge is reward: Learning optimal exploration by predictive reward cashing

Re-assign community