Towards Human-Agent Communication via the Information Bottleneck Principle

  title={Towards Human-Agent Communication via the Information Bottleneck Principle},
  author={Mycal Tucker and Julie A. Shah and Roger Philip Levy and Noga Zaslavsky},
—Emergent communication research often focuses on optimizing task-specific utility as a driver for communication. However, human languages appear to evolve under pressure to efficiently compress meanings into communication signals by optimizing the Information Bottleneck tradeoff between informa- tiveness and complexity. In this work, we study how trading off these three factors — utility, informativeness, and complexity — shapes emergent communication, including compared to human communication… 

Figures from this paper



Exploring Zero-Shot Emergent Communication in Embodied Multi-Agent Populations

This work shows that under realistic assumptions, a non-uniform distribution of intents and a common-knowledge energy cost, agents that learn to communicate via actuating their joints in a 3D environment can find protocols that generalize to novel partners.

Emergent Discrete Communication in Semantic Spaces

Inspired by word embedding techniques from natural language processing, this work proposes neural agent architectures that enables them to communicate via discrete tokens derived from a learned, continuous space that optimizes communication over a wide range of scenarios, whereas one-hot tokens are only optimal under restrictive assumptions.

Quasi-Equivalence Discovery for Zero-Shot Emergent Communication

This work presents a novel problem setting and the Quasi-Equivalence Discovery (QED) algorithm that allows for zeroshot coordination (ZSC) and shows that these two factors lead to unique optimal ZSC policies in referential games, where agents use the energy cost of the messages to communicate intent.

Anti-efficient encoding in emergent communication

Surprisingly, networks develop an anti-efficient encoding scheme, in which the most frequent inputs are associated to the longest messages, and messages in general are skewed towards the maximum length threshold.

Learning Efficient Multi-agent Communication: An Information Bottleneck Approach

It is proved that the limited bandwidth constraint requires low-entropy messages throughout the transmission and inspired by the information bottleneck principle, a valuable and compact communication protocol and a weight-based scheduler are learned.

Biases for Emergent Communication in Multi-agent Reinforcement Learning

This work introduces inductive biases for positive signalling and positive listening, which ease the learning problem in emergent communication and applies these methods to a more extended environment, showing that agents with these inductive bias achieve better performance.

Multi-Agent Cooperation and the Emergence of (Natural) Language

It is shown that two networks with simple configurations are able to learn to coordinate in the referential game and how to make changes to the game environment to cause the "word meanings" induced in the game to better reflect intuitive semantic properties of the images.

Social Influence as Intrinsic Motivation for Multi-Agent Deep Reinforcement Learning

Empirical results demonstrate that influence leads to enhanced coordination and communication in challenging social dilemma environments, dramatically increasing the learning curves of the deep RL agents, and leading to more meaningful learned communication protocols.

Learning to Communicate with Deep Multi-Agent Reinforcement Learning

By embracing deep neural networks, this work is able to demonstrate end-to-end learning of protocols in complex environments inspired by communication riddles and multi-agent computer vision problems with partial observability.

A reinforcement-learning approach to efficient communication

It is suggested that RL offers a powerful and flexible computational framework that can contribute to the development of communication schemes for color names that are near-optimal in an information-theoretic sense and may shape color-naming systems across languages.