Emergence of Grounded Compositional Language in Multi-Agent Populations

TL;DR

This paper explores how grounded, compositional language can naturally emerge among multi-agent systems to facilitate goal-oriented communication, including non-verbal cues, through a novel learning environment.

Contribution

It introduces a multi-agent environment and methods demonstrating the emergence of structured, grounded language and non-verbal communication.

Findings

Emergence of a structured, compositional language with vocabulary and syntax

Development of non-verbal communication like pointing and guiding

Language facilitates goal achievement in multi-agent interactions

Abstract

By capturing statistical patterns in large corpora, machine learning has enabled significant advances in natural language processing, including in machine translation, question answering, and sentiment analysis. However, for agents to intelligently interact with humans, simply capturing the statistical patterns is insufficient. In this paper we investigate if, and how, grounded compositional language can emerge as a means to achieve goals in multi-agent populations. Towards this end, we propose a multi-agent learning environment and learning methods that bring about emergence of a basic compositional language. This language is represented as streams of abstract discrete symbols uttered by agents over time, but nonetheless has a coherent structure that possesses a defined vocabulary and syntax. We also observe emergence of non-verbal communication such as pointing and guiding when language communication is unavailable.