Closing the Intent-to-Behavior Gap via Fulfillment Priority Logic
This work addresses a fundamental problem for practitioners in reinforcement learning by providing a novel method to define and prioritize objectives, though it is incremental as it builds on existing multi-objective reinforcement learning approaches.
The paper tackles the challenge of translating intended behavioral objectives into reward functions in reinforcement learning, particularly for multi-objective scenarios like robotics where performance conflicts with energy conservation, and introduces Fulfillment Priority Logic (FPL) with a Balanced Policy Gradient algorithm that achieves up to 500% better sample efficiency compared to Soft Actor Critic.
Practitioners designing reinforcement learning policies face a fundamental challenge: translating intended behavioral objectives into representative reward functions. This challenge stems from behavioral intent requiring simultaneous achievement of multiple competing objectives, typically addressed through labor-intensive linear reward composition that yields brittle results. Consider the ubiquitous robotics scenario where performance maximization directly conflicts with energy conservation. Such competitive dynamics are resistant to simple linear reward combinations. In this paper, we present the concept of objective fulfillment upon which we build Fulfillment Priority Logic (FPL). FPL allows practitioners to define logical formula representing their intentions and priorities within multi-objective reinforcement learning. Our novel Balanced Policy Gradient algorithm leverages FPL specifications to achieve up to 500\% better sample efficiency compared to Soft Actor Critic. Notably, this work constitutes the first implementation of non-linear utility scalarization design, specifically for continuous control problems.