How to Exhibit More Predictable Behaviors
This work addresses predictability issues for agents in interactive systems, but it appears incremental as it builds on existing observer-aware MDP frameworks.
The paper tackles the problem of enabling agents to exhibit more predictable behaviors by optimizing for an external observer's predictions, addressing uncertainties in environment dynamics and agent policies. It proposes performance criteria through reward functions based on observer beliefs, shows these can be represented by standard MDPs, and analyzes properties theoretically and empirically on grid-world problems.
This paper looks at predictability problems, i.e., wherein an agent must choose its strategy in order to optimize the predictions that an external observer could make. We address these problems while taking into account uncertainties on the environment dynamics and on the observed agent's policy. To that end, we assume that the observer 1. seeks to predict the agent's future action or state at each time step, and 2. models the agent using a stochastic policy computed from a known underlying problem, and we leverage on the framework of observer-aware Markov decision processes (OAMDPs). We propose action and state predictability performance criteria through reward functions built on the observer's belief about the agent policy; show that these induced predictable OAMDPs can be represented by goal-oriented or discounted MDPs; and analyze the properties of the proposed reward functions both theoretically and empirically on two types of grid-world problems.