Learning Compositional Negation in Populations of Roth-Erev and Neural Agents

TL;DR

This paper investigates how populations of neural and Roth-Erev agents learn simple compositional negation rules in signaling games, revealing the robustness of such properties across various models and agent types.

Contribution

It introduces a novel signaling game variant for studying negation learnability, extends analysis to populations, and compares neural and Roth-Erev agents under different conditions.

Findings

Basic compositional negation is robustly learnable.

Population size influences the stability of signaling systems.

Neural agents effectively learn negation with reinforcement learning.

Abstract

Agent-based models and signalling games are useful tools with which to study the emergence of linguistic communication in a tractable setting. These techniques have been used to study the compositional property of natural languages, but have been limited in how closely they model real communicators. In this work, we present a novel variant of the classic signalling game that explores the learnability of simple compositional rules concerning negation. The approach builds on the work of Steinert-Threlkeld (2016) by allowing agents to determine the identity of the "function word" representing negation while simultaneously learning to assign meanings to atomic symbols. We extend the analysis with the introduction of a population of concurrently communicating agents, and explore how the complications brought about by a larger population size affect the type and stability of the signalling systems learned. We also relax assumptions of the parametric form of the learning agents and examine how neural network-based agents optimized through reinforcement learning behave under various task settings. We find that basic compositional properties are robustly learnable across a wide range of model relaxations and agent instantiations.