Artifacts as Memory Beyond the Agent Boundary

TL;DR

This paper formalizes how environmental resources, termed artifacts, can serve as external memory in reinforcement learning, reducing internal memory needs and enabling more efficient policy learning.

Contribution

It introduces a mathematical framework for environmental memory in RL and demonstrates its effectiveness through experiments with spatial path observations.

Findings

Artifacts can reduce the memory required for RL agents.

Environmental observations implicitly serve as external memory.

Memory reduction occurs naturally through sensory streams.

Abstract

The situated view of cognition holds that intelligent behavior depends not only on internal memory, but on an agent's active use of environmental resources. Here, we begin formalizing this intuition within Reinforcement Learning (RL). We introduce a mathematical framing for how the environment can functionally serve as an agent's memory, and prove that certain observations, which we call artifacts, can reduce the information needed to represent history. We corroborate our theory with experiments showing that when agents observe spatial paths, the amount of memory required to learn a performant policy is reduced. Interestingly, this effect arises unintentionally, and implicitly through the agent's sensory stream. We discuss the implications of our findings, and show they satisfy qualitative properties previously used to ground accounts of external memory. Moving forward, we anticipate further work on this subject could reveal principled ways to exploit the environment as a substitute for explicit internal memory.