WideningthePipelineinHuman-Guided ReinforcementLearningwithExplanationand Context-AwareDataAugmentation