Learning Who Talks to Whom: Role-Aware Communication Topology Scheduling for Efficient LLM Multi-Agent Collaboration

Haibin Li; Wenya Yu; Xuefeng Jiang; Jia Li; Xiao Han; Shuai Su; Wei Wei

cs.MAOct 19, 2026

Learning Who Talks to Whom: Role-Aware Communication Topology Scheduling for Efficient LLM Multi-Agent Collaboration

Authors: Haibin Li, Wenya Yu, Xuefeng Jiang, Jia Li, Xiao Han, Shuai Su, Wei Wei

Organizations: Beihang University · Nanyang Technological University · University of Chinese Academy of Sciences

Abstract

Multi-agent systems driven by large language models (LLMs) rely on structured communication topologies to facilitate information exchange and collaborative reasoning. However, effectively scheduling these topologies to dynamically determine who talks to whom and when to communicate remains a fundamental challenge in balancing collective reasoning performance against exponential token overhead. Existing topology scheduling approaches either enforce static communication graphs that cannot adapt to dynamic problem-solving states, or overlook agent role complementarity while failing to explicitly incorporate the synthesized graph structure into team-level decision evaluation. To tackle these limitations, we propose ACTS, a role-aware adaptive communication topology scheduling framework for efficient LLM multi-agent collaboration. Orchestrated within a Centralized Training with Decentralized Execution (CTDE) paradigm, ACTS introduces a novel dual-attention mechanism. Specifically, a Role-conditioned Relational Attention module dynamically selects variable-cardinality interaction partners based on agent roles and evolving execution states, while a Topological Attention Mixer explicitly embeds the directed communication topology into team value decomposition for precise credit assignment. Comprehensive evaluations across 9 challenging benchmarks covering code generation, multi-step reasoning, general knowledge, and competitive mathematics demonstrate that ACTS consistently outperforms state-of-the-art baselines, boosting average accuracy by up to 6.19 percentage points while reducing token consumption by over 15%. The project website is available at https://acts-mas.github.io/ACTS/.

Figures & tables

Figure 1: Motivation and method comparison. (a) Topology scheduling for efficient multi-agent collaboration. (b) Method comparison highlighting ACTS ’s role-aware partner selection and topology-aware value mixing.

Figure 2: Overview of ACTS . (a) Decentralized execution. Local action-value functions select communication modes, while role-aware relational attention selects directed partners and induces the adaptive communication topology. (b) Centralized training. Episode transitions are stored in a replay buffer. Topological attention and the mixing network combine the realized topology $G^{t}$ with local utilities to estimate $Q_{\mathrm{tot}}^{t}$ , which is optimized by TD learning under a task-accuracy and token-cost objective.

Figure 3: Architecture of ACTS . Role-aware relational attention guides directed partner selection. Topological attention weights local utilities according to directed neighborhood representations, and the mixing network constructs $Q_{\mathrm{tot}}^{t}$ . These modules are jointly optimized by minimizing the TD loss within a value function factorization MARL framework under CTDE.

Method	LiveCode Bench	Human Eval	HLE	Hotpot QA	MMLU- Pro	AIME 2025	AIME 2026	Beyond- AIME	HMMT 2025	Avg.
LLM backbone: GPT-OSS-120B
Direct Prompting	88.57	89.63	42.40	20.00	70.00	60.00	86.67	16.00	46.67	57.77
Static topology
LLM-Debate	92.00	88.41	42.80	21.35	71.43	63.33	86.67	18.00	46.67	58.96
AutoGen	95.43	90.24	41.20	32.03	77.14	70.00	80.00	18.00	46.67	61.19
LangGraph	93.14	89.02	42.40	40.00	84.29	76.67	86.67	23.00	50.00	65.02

Table 1: Accuracy (%) across nine benchmarks. Bold marks the best result in each column; ACTS is highlighted in light blue. Avg. is the unweighted macro-average over the nine benchmarks.

Figure 4: Accuracy vs. token usage for all methods on two representative tasks (HotpotQA with $\mathrm{IG}>0$ and HLE with $\mathrm{IG}\approx 0$ ) and two backbone models. The dashed line traces the Pareto frontier of methods not dominated on both accuracy and cost. ACTS lies on the frontier in all four panels: no other multi-agent method reaches its accuracy at lower or equal cost, and only single-agent prompting is cheaper, at a substantial accuracy loss.

Figure 5: Mean $\Delta$ ACC of each topology family over single-agent prompting, by backbone model. The dashed line partitions benchmarks into $\mathrm{IG}>0$ on the left and $\mathrm{IG}\approx 0$ within $\pm 2/n$ on the right.

Figure 6: Action distribution by task on DeepSeek-V4-Flash. On HotpotQA, the policy sustains agent activity and broadcast communication; on HLE, it deactivates a substantial fraction of rounds and relies on pull-based retrieval.

	GPT-OSS-120B		DeepSeek-V4-Flash
Variant	$\overline{\mathrm{ACC}}$	$\overline{\mathrm{Token}}$	$\overline{\mathrm{ACC}}$	$\overline{\mathrm{Token}}$
ACTS	75.01	5.97	69.43	4.03
(a) Topological mixer replaced by QMIX	71.83	7.41	63.24	4.96
(b) Learned routing replaced by fixed partners	72.50	8.10	66.10	6.20
(c) Role embeddings removed	73.40	6.80	67.10	4.60

Table 2: Ablation results on the two backbone models.

\overline{\mathrm{ACC}}

and

\overline{\mathrm{Token}}

denote the macro-average accuracy (%) and macro-average per-sample token usage (

\times 10^{3}

) across the nine benchmarks. The full ACTS model is highlighted.

Appendix figures & tables5 assets

Supplementary material from the paper’s appendix.

Appendix

y_{k}=\begin{cases}r_{k},&d_{k}=1,\\ r_{k}+\gamma Q_{\mathrm{tot}}^{-}(\bm{a}_{k}^{*,+};\mathcal{C}_{k}^{+}),&d_{k}=0,\end{cases}

Appendix

Algorithm 1 ACTS Centralized Training

Benchmark	Domain	$n$	$x=2/n$	IG (GPT-OSS-120B)	IG (DeepSeek-V4-Flash)
LiveCodeBench	Code	175	0.0114	$>0$	$>0$
HumanEval	Code	164	0.0122	$\approx 0$	$\approx 0$
HLE	Reasoning	250	0.0080	$\approx 0$	$\approx 0$
HotpotQA	Multi-hop QA	740	0.0027	$>0$	$>0$
MMLU-Pro	Reasoning	70	0.0286	$>0$	$>0$
AIME 2025	Math	30	0.0667	$>0$	$>0$

Appendix

Table 3: Per-benchmark evaluation samples, IG thresholds, and classifications under GPT-OSS-120B and DeepSeek-V4-Flash.

Figure 7: Distribution of the six actions across benchmarks under DeepSeek-V4-Flash. Each cell gives the proportion of rounds in which the corresponding action is selected. The learned policy favours broadcast and attn-broadcast on benchmarks where the agents contribute complementary information, and shifts toward off and attn-receive where a single perspective suffices.

Figure 8: Trajectory-level case study on one HotpotQA question and one HLE question. On HotpotQA the connected topology contracts into a sparse pull while one agent retains the solo action as an independent reference. On HLE one agent settles the answer under the solo action, a single attn-broadcast check is routed to the agent that requires it, and the remaining agents are deactivated through the off action, triggering early termination under the majority-inactive condition.

	HotpotQA		HLE
Method	GPT-OSS-120B	DeepSeek-V4-Flash	GPT-OSS-120B	DeepSeek-V4-Flash
Direct Prompting	5.8	3.6	4.5	2.8
LLM-Debate	14.1	9.5	12.3	8.2
AutoGen	16.0	10.8	13.8	9.3
LangGraph	16.4	11.0	14.0	9.5
Agent Framework	16.7	11.3	14.2	9.7

Appendix

Table 4: Average per-sample token usage (

\times 10^{3}

) on HotpotQA, a high-information-gain (

\mathrm{IG}>0

) task, and HLE, a low-information-gain (

\mathrm{IG}\approx 0

) task, for both backbone models. ACTS (highlighted) consumes the fewest tokens among all multi-agent methods in every column; only single-agent prompting uses fewer, at a large accuracy cost.

Multi-Agent Communication Topology

See all 34 papers

GoAgent: Group-of-Agents Communication Topology Generation for LLM-based Multi-Agent Systems
Mar 20, 2026Hongjiang Chen, Xin Zheng, Yixin Liu +7Multi-Agent Communication Topology Multi-Agent Communication
School of Cyberspace, Hangzhou Dianzi University, Hangzhou, China · School of Computing Technologies, RMIT University, Melbourne, Australia · School of Information and Communication Technology, Griffith University, Goldcoast, Australia
Agent Q-Mix: Selecting the Right Action for LLM Multi-Agent Systems through Reinforcement Learning
Apr 1, 2026Eric Hanchen Jiang, Levina Li, Rui Sun +9Multi-Agent Communication Topology Multi-Agent Reasoning
University of California Los Angeles · University of Washington · Northwestern University
Learning Transferable Topology Priors for Multi-Agent LLM Collaboration Across Domains
May 17, 2026Taolin Zhang, Zijie Zhou, Jiuheng Wan +4Multi-Agent LLM Systems Multi-Agent System Optimization
Hefei University of Technology, Hefei 230002, China · China University of Petroleum (Beijing), Beijing 102249, China · East China Normal University, Shanghai 200062, China +1

Mar 20, 2026J/K move · Enter open · S save

GoAgent: Group-of-Agents Communication Topology Generation for LLM-based Multi-Agent Systems

Hongjiang Chen, Xin Zheng, Yixin Liu, Pengfei Jiao+6

School of Cyberspace, Hangzhou Dianzi University, Hangzhou, China · School of Computing Technologies, RMIT University, Melbourne, Australia · School of Information and Communication Technology, Griffith University, Goldcoast, Australia

Multi-Agent Communication Topology Multi-Agent Communication Multi-Agent LLMs

Open paper AskPDF arXiv

Large language model (LLM)-based multi-agent systems (MAS) have demonstrated exceptional capabilities in solving complex tasks, yet their effectiveness depends heavily on the underlying communication topology that coordinates agent interactions. Within these systems, successful problem-solving often necessitates task-specific group structures to divide and conquer subtasks. However, most existing approaches generate communication topologies in a node-centric manner, leaving group structures to emerge implicitly from local connectivity decisions rather than modeling them explicitly, often leading to suboptimal coordination and unnecessary communication overhead. To address this limitation, we propose GoAgent (Group-of-Agents), a communication topology generation method that explicitly treats collaborative groups as the atomic units of MAS construction. Specifically, GoAgent first enumerates task-relevant candidate groups through an LLM and then autoregressively selects and connects these groups as atomic units to construct the final communication graph, jointly capturing intra-group cohesion and inter-group coordination. To mitigate communication redundancy and noise propagation inherent in expanding topologies, we further introduce a conditional information bottleneck (CIB) objective that compresses inter-group communication, preserving task-relevant signals while filtering out redundant historical noise. Extensive experiments on six benchmarks demonstrate the state-of-the-art performance of GoAgent with 93.84% average accuracy while reducing token consumption by about 17%.

Abstract

Multi-Agent Communication Topology

Explore similar work

GoAgent: Group-of-Agents Communication Topology Generation for LLM-based Multi-Agent Systems

Agent Q-Mix: Selecting the Right Action for LLM Multi-Agent Systems through Reinforcement Learning

Learning Transferable Topology Priors for Multi-Agent LLM Collaboration Across Domains