AI Paper Atlas

Research map / Multi-agent systems and games / LLM agents

Navigating the Energy Landscape of Collaboration: Multi-Agent Communication Graph Generation via Score-Based Diffusion

GuanHao Zhao, Wenbo Lu, Cheng Cheng, Zhenya Huang, Wei Song, Zhiding Liu, Runze Wu, Enhong Chen

ICML 2026 · Multi-agent systems and games › LLM agents · approach: LLM multi-agent decision making

Read the paperCheck your idea against 137,217 papers

AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Proposes generating task-adaptive communication graphs for LLM-based multi-agent systems using score-based diffusion, framed through an energy landscape view of collaboration.

Problem

Discovering task-adaptive communication graph structures for LLM-based multi-agent systems within a combinatorial search space, which existing heuristic pruning and autoregressive generation methods handle without a unified theoretical framework.

Approach

Treats the communication graph as a sample from an energy landscape and generates graphs with a score-based diffusion model, offering a unified theoretical framing for self-organizing agent topologies (only the abstract was available; details beyond this are not stated).

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add joint graph-node-edge score modeling (vs w/o Joint) (ablation)MAGE · MATHaccuracy (%): 70.09 -> 72.22Improved
add joint modeling (ablation)MAGE · GSM8Kaccuracy (%): 92.27 -> 94.35Improved
add joint modeling (ablation)MAGE · MBPPaccuracy (%): 83.75 -> 84.57Improved
add energy-guided sampling (vs w/o Guidance) (ablation)MAGE · GSM8Kaccuracy (%): 93.85 -> 94.35Improved
add energy-guided sampling (ablation)MAGE · MBPPaccuracy (%): 83.78 -> 84.57Improved
add energy-guided sampling (ablation)MAGE · MATHaccuracy (%): 71.43 -> 72.22Improved
first-order energy gradient guidance instead of zeroth-order random perturbation guidance (ablation)MAGE · MATH, 20 diffusion stepsaccuracy (%): 69.19 -> 74.33Improved
first-order guidance vs zeroth-order guidance across diffusion steps (ablation)MAGE · MATHaccuracy (%): MAGE-First 72.22 vs MAGE-Zero 70.31 at 50 stepsImproved
increase guidance scale eta from 1.0 to 5.0 (ablation)MAGE · MBPPaccuracy (%): 84.31% to 87.23% fluctuationNo clear effect
MAGE graphs versus GTD on token costMAGE · GSM8Ktotal cost ($): 1.48 -> 1.43Improved
ADAS LLM-driven workflow generationADAS · average over 6 benchmarksaccuracy (%): 81.68 -> 80.19Hurt
MAGE versus best prior baseline GTDMAGE · MATHaccuracy (%): 68.44 -> 72.22Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy