Proposing a generic early stopping method for direct policy search that significantly reduces computation time without the need for problem-specific knowledge.
