# Communicating Plans, Not Percepts: Scalable Multi-Agent Coordination with Embodied World Models

arXiv: https://arxiv.org/abs/2508.02912 · HTML: https://mlsanity.com/paper/2508.02912/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Placed in World models for language agents, one of 11 directions in World models.
- Builds on 19 of its 29 references that are in the map.
- First posted on arXiv 4 Aug 2025.

## Placed in

- World models / [World models for language agents](https://mlsanity.com/wm/world-models-for-language-agents/)

## Builds on (most cited first, 19 of 19)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Dream to Control: Learning Behaviors by Latent Imagination](https://mlsanity.com/paper/1912.01603/) (2019)
- [Multi-Agent Actor-Critic for Mixed Cooperative-Competitive Environments](https://mlsanity.com/paper/1706.02275/) (2017)
- [Mastering Diverse Domains through World Models](https://mlsanity.com/paper/2301.04104/) (2023)
- [Mastering Atari with Discrete World Models](https://mlsanity.com/paper/2010.02193/) (2020)
- [The StarCraft Multi-Agent Challenge](https://mlsanity.com/paper/1902.04043/) (2019)
- [Learning to Communicate with Deep Multi-Agent Reinforcement Learning](https://mlsanity.com/paper/1605.06676/) (2016)
- [Leveraging Procedural Generation to Benchmark Reinforcement Learning](https://mlsanity.com/paper/1912.01588/) (2019)
- [Learning Multiagent Communication with Backpropagation](https://mlsanity.com/paper/1605.07736/) (2016)
- [On the Utility of Learning about Humans for Human-AI Coordination](https://mlsanity.com/paper/1910.05789/) (2019)
- [Emergence of Grounded Compositional Language in Multi-Agent Populations](https://mlsanity.com/paper/1703.04908/) (2017)
- [Google Research Football: A Novel Reinforcement Learning Environment](https://mlsanity.com/paper/1907.11180/) (2019)
- [The Complexity of Decentralized Control of Markov Decision Processes](https://mlsanity.com/paper/1301.3836/) (2013)
- [PettingZoo: Gym for Multi-Agent Reinforcement Learning](https://mlsanity.com/paper/2009.14471/) (2020)
- [A Review of Cooperative Multi-Agent Deep Reinforcement Learning](https://mlsanity.com/paper/1908.03963/) (2019)
- [Modeling Others using Oneself in Multi-Agent Reinforcement Learning](https://mlsanity.com/paper/1802.09640/) (2018)
- [Towards Collaborative Intelligence: Propagating Intentions and Reasoning for Multi-Agent Coordination with Large Language Models](https://mlsanity.com/paper/2407.12532/) (2024)
- [Imagine, Initialize, and Explore: An Effective Exploration Method in Multi-Agent Reinforcement Learning](https://mlsanity.com/paper/2402.17978/) (2024)
- [Multi-Agent Intention Sharing via Leader-Follower Forest](https://mlsanity.com/paper/2112.01078/) (2021)

## Built on by (most cited first, 3 of 3)

- [Mosaic: Runtime-Efficient Multi-Agent Embodied Planning](https://mlsanity.com/paper/2607.09603/) (2026)
- [When to Communicate: Belief Distributions and KL Divergence for Principled Gating in Multi-Agent RL](https://mlsanity.com/paper/2608.14559/) (2026)
- [Emergence, Not Bandwidth: Physical Coupling and the Limits of Learned Multi-Agent Communication](https://mlsanity.com/paper/2609.34373/) (2026)

## Abstract (arXiv)

Robust coordination is critical for effective decision-making in multi-agent systems, especially under partial observability. A central question in Multi-Agent Reinforcement Learning (MARL) is whether to engineer communication protocols or learn them end-to-end. We investigate this dichotomy using embodied world models. We propose and compare two communication strategies for a cooperative task-allocation problem. The first, Learned Direct Communication (LDC), learns a protocol end-to-end. The second, Intention Communication, uses an engineered inductive bias: a compact, learned world model, the Imagined Trajectory Generation Module (ITGM), which uses the agent's own policy to simulate future states. A Message Generation Network (MGN) then compresses this plan into a message. We evaluate these approaches on goal-directed interaction in a grid world, a canonical abstraction for embodied AI problems, while scaling environmental complexity. Our experiments reveal that while emergent communication is viable in simple settings, the engineered, world model-based approach shows superior performance, sample efficiency, and scalability as complexity increases. These findings advocate for integrating structured, predictive models into MARL agents to enable active, goal-driven coordination.