SupplementaryMaterialforRethinkingValue FunctionLearningforGeneralizationin ReinforcementLearning