Metrics
1 Record Views
Journal article
GEMA: Graph Embeddings for Multi‐Agent Coordination
The AI magazine, Vol.47(3), e70088
09/2026
Web of Science ID: WOS:001856320300001
Abstract
Many cooperative multi‐agent tasks are naturally defined by graph‐structured objectives, where agents must collectively reach, for example, a desired relational configuration or satisfy a set of constraints. These objectives often encode spatial arrangements, inter‐agent relations, or constraints that can be formalized as target graphs. However, current goal‐conditioned multi‐agent reinforcement learning (MARL) algorithms do not employ these symbolic and structured representations to direct their agents towards effective strategies. We propose Graph Embeddings for Multi‐Agent Coordination (GEMA), which couples any cooperative learner with a State‐Goal Graph Encoder (SGE). The SGE is pre‐trained in a contrastive manner to embed state graphs in a common metric space. At run time, each agent builds the state graph, queries the SGE, and computes a scalar distance to the goal embedding. The calculated distance can then be used as an intrinsic reward signal for the agents and incorporated into each agent's observation, providing feedback on the task progress. Experiments on challenging benchmarks for centralized training with decentralized execution (CTDE), including cooperative navigation, load balancing, and the StarCraft Multi‐Agent Challenge (v2), demonstrate that GEMA accelerates convergence and improves team returns, outperforming both standard and objective‐driven state‐of‐the‐art MARL baselines. Code repository: .
Files and links (2)
Related links
Details
- Title
- GEMA: Graph Embeddings for Multi‐Agent Coordination
- Publication Details
- The AI magazine, Vol.47(3), e70088
- Resource Type
- Journal article
- Publisher
- John Wiley & Sons, Inc
- Number of pages
- 14
- Grant note
- Air Force Research Laboratory: FA8651-25-1-0003
This material is based on research sponsored by Air Force Research Laboratory under agreement number FA8651-25-1-0003. The U.S. Government is authorized to reproduce and distribute reprints for Governmental purposes, notwithstanding any copyright notation thereon. Approved for public release: distribution unlimited.
- Copyright
- © 2026 The Author(s).
- Identifiers
- WOS:001856320300001; 99381871230006600
- Academic Unit
- Institute for Human and Machine Cognition; Intelligent Systems and Robotics; Hal Marcus College of Science and Engineering
- Language
- English