Skip to content

MA-WAM: Multi-Agent World-Action Model for Test-Time Planning

Sep 2026 · 1 citation · 49 references
Computer Science

TL;DR

This work proposes Multi-Agent World-Action Model (MA-WAM), a test-time planning framework that enables a frozen multi-agent flow policy to evaluate futures of candidate joint actions and is the first test-time world-model planner for multi-agent flow policies.

Abstract

Multi-agent cooperative tasks require different agents to execute a joint action simultaneously, and each agent's action affects both the observations and responses of the other agents. Hence, a world model is needed to predict the team return resulting from the joint actions of all agents. A naive extension directly applies a single-agent world model to each agent's action when predicting the team return step by step. However, such an extension fails to capture the dependencies among the simultaneous actions of multiple agents. We propose Multi-Agent World-Action Model (MA-WAM), a test-time planning framework that enables a frozen multi-agent flow policy to evaluate futures of candidate joint actions. To our knowledge, MA-WAM is the first test-time world-model planner for multi-agent flow policies. MA-WAM predicts the consequences of each joint action according to cross-agent dependencies and enables efficient candidate scoring. Across 30 offline multi-agent reinforcement learning (MARL) settings on MAMuJoCo, SMAC, and MPE, MA-WAM achieves mean relative gains of 22.0% over direct execution and 25.6% over uniform action selection. Under the standard evaluation protocol on an A100 GPU, MA-WAM adds 12.1 ms, accounting for 2.5% of the measured generation-and-scoring time.

View source

Similar papers

Preprint Oct 2026

Self-Improving Agents with World Model Predictive Paths

A test-time framework that combines multi-branch predictive analysis with experience-based adaptation, inspired by Model Predictive Path Integral (MPPI) control, and demonstrates the strong overall performance and the effectiveness of experience-guided predictive control.

Hao-Chen Lu, Tian-Yu Gao, Yaoyao Liu et al. · 0 citations
#artificial intelligence Preprint Oct 2026

CB-Agent: Concept-Based In-Context Learning for Sequential Action Selection

In-context learning (ICL) allows large language models to use demonstrations without updating their parameters. For interactive agents, demonstrations are trajectories of reasoning, actions, and observations that illustrate how to complete a task. Reusing demonstrations in new situations is challenging: directly inputt...

Qing-Yue Zhang, Hua-Ze Tang, Min-Yu Zhang et al. · 0 citations
Preprint Oct 2026

From Single-Actor Environments to Multi-Actor Event Worlds: Can LLMs Be Event World Models?

World models help agents assess the consequences of their actions before acting. Many real systems, however, involve multiple actors who make their own decisions, can evolve even when the agent takes no action, and are only partially observed through event logs. We formalize these systems as event worlds and ask whethe...

Jun-Tong Ni, Ze-Wen Liu, Bo-Han Wang et al. · 0 citations
Sep 2026

AgentWorld: Benchmarking Long-Horizon Collaboration of Multi-agent LLMs

To quantify collaboration effectiveness in addition to conventional binary task success, Causal Collaboration Effectiveness (CCE) is proposed, a graph-based metric that traces causal dependencies between agent actions and measures what fraction of a team's effort actually contributed to the outcome.

Raphael Shu, Yu-Sen Zhang, Y. Cho et al. · 2 citations
Preprint Oct 2026

LAW: Unifying Language, Action, and World Modeling for Agentic Physical Intelligence

LAW, a closed-loop VLM--WAM framework that unifies language reasoning, action generation, and world modeling through semantic subtask progress, is presented, establishing a new open-source state of the art on RoboTwin.

Jia-Jun Cao, Kai Tang, Bo-Hua Zhang et al. · 0 citations
#artificial intelligence Preprint Oct 2026

Learning to Steer LLM Agents and Multi-Agent Systems

This work proposes a steering agent that supervises execution on the fly and steers the system back on track when drift occurs, and jointly train the two paths with reinforcement learning, combining a pairwise ranking objective with group-relative policy optimization (GRPO).

Le-Yang Shen, Ya-Xin Gong, Yang Zhang et al. · 0 citations

Related blog posts

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.