# Perspective Taking in Deep Reinforcement Learning Agents

## TL;DR

This paper explores developing artificial agents capable of perspective taking, inspired by animal behavior, and demonstrates that neural network-based agents can learn simple perspective-taking tasks through reinforcement learning.

## Contribution

It introduces a perspective taking task for artificial agents and investigates how different information encoding affects learning, advancing AI's social interaction capabilities.

## Key findings

- Agents can learn perspective taking via reinforcement learning.
- Encoding information in allocentric or egocentric forms influences learning efficiency.
- Progress towards human-like social AI interactions.

## Abstract

Perspective taking is the ability to take the point of view of another agent. This skill is not unique to humans as it is also displayed by other animals like chimpanzees. It is an essential ability for social interactions, including efficient cooperation, competition, and communication. Here we present our progress toward building artificial agents with such abilities. We implemented a perspective taking task inspired by experiments done with chimpanzees. We show that agents controlled by artificial neural networks can learn via reinforcement learning to pass simple tests that require perspective taking capabilities. We studied whether this ability is more readily learned by agents with information encoded in allocentric or egocentric form for both their visual perception and motor actions. We believe that, in the long run, building better artificial agents with perspective taking ability can help us develop artificial intelligence that is more human-like and easier to communicate with.