Essay.09-Utility_Representations-HulingxiaoHe-2301112059

Hulingxiao He

Keywords: Utility, Reward

Abstract: The concept of utility is commonly employed to quantify worth and value, closely tied to human preferences and capable of offering insights into human behaviors. Utility functions have been crafted by researchers to articulate decision-making challenges, particularly within the domain of reinforcement learning. This essay aims to explore the distinctions between utility functions and reward functions, delving into the representation of utility in the context of reinforcement learning. Finally, it emphasizes the future representation of utility through curiosity-driven approaches and the like.

Submission Number: 164