On The Expressivity Of Markov Reward

Abstract

Reward is the driving force for reinforcement-learning agents. This paper is dedicated to understanding the expressivity of reward as a way to capture tasks that we would want an agent to perform. We frame this study around three new abstract notions of "task" that might be desirable: (1) a set of acceptable behaviors, (2) a partial ordering over behaviors, or (3) a partial ordering over trajectories. Our main results prove that while reward can express many of these tasks, there exist instances of each task type that no Markov reward function can capture. We then provide a set of polynomial-time algorithms that construct a Markov reward function that allows an agent to optimize tasks of each of these three types, and correctly determine when no such reward function exists. We conclude with an empirical study that corroborates and illustrates our theoretical findings.

Stats

• citations0
• S2 citations—
• github stars0
• HF likes0
• heat score0.00
• arxiv keyabel2021on

Related papers

• Informativeness Of Reward Functions In Reinforcement Learning (2024)2.26
• Goodhart's Law In Reinforcement Learning (2023)0.00
• Reward Is Enough For Convex Mdps (2021)0.00
• Tiered Reward: Designing Rewards For Specification And Fast Learning Of Desired Behavior (2022)0.00
• Scalable Agent Alignment Via Reward Modeling: A Research Direction (2018)0.00
• Learning Non-markovian Reward Models In Mdps (2020)0.00
• Unifying Task Specification In Reinforcement Learning (2016)0.00
• Extended Markov Games To Learn Multiple Tasks In Multi-agent Reinforcement Learning (2020)3.58