Essay.08-ChenhaoZhou-2100017709
Keywords: Utility; Reinforcement learning
Abstract: Utility theory, originally served as a definition for modeling human decisionmaking process, has long been considered an internal estimate of human choices. The reward signal as an effective assessment of human utility is widely applied in reinforcement learning. In this essay, we will divide the construction of reward signals into two types to discuss: explicit and implicit. Additionally, primarily in the context of reward signals, there are surely a number of debate about whether it fully represents the human utility.
Submission Number: 162
Loading