Learning Reward Functions from a Combination of Demonstration and Evaluative Feedback | IEEE Conference Publication | IEEE Xplore