A Machine With Human-Like Memory Systems

TL;DR

This paper presents a cognitive-inspired agent with semantic and episodic memory systems, demonstrating improved performance in a custom environment and highlighting benefits of human-machine collaboration.

Contribution

It introduces a dual-memory agent model inspired by cognitive science and a new environment for testing memory-based learning and collaboration.

Findings

Agents with both memory systems outperform single-memory agents.

Collaboration between machine and human agents enhances performance.

The 'Room' environment effectively tests memory encoding and retrieval.

Abstract

Inspired by the cognitive science theory, we explicitly model an agent with both semantic and episodic memory systems, and show that it is better than having just one of the two memory systems. In order to show this, we have designed and released our own challenging environment, "the Room", compatible with OpenAI Gym, where an agent has to properly learn how to encode, store, and retrieve memories to maximize its rewards. The Room environment allows for a hybrid intelligence setup where machines and humans can collaborate. We show that two agents collaborating with each other results in better performance than one agent acting alone.