Elamur Research⌕
☰
home / papers / 2026 / EmoMAS: Emotion-Aware Multi-Agent System for High-Stakes Edge-Deployable Negotiation with Bayesian Orchestration

EmoMAS: Emotion-Aware Multi-Agent System for High-Stakes Edge-Deployable Negotiation with Bayesian Orchestration

Paper · 2026 · Yunbo Long, Yuhan Liu and Liming Xu

Summary

Negotiation agents built on cloud LLMs are costly and expose sensitive data. Small language models (SLMs) could run locally on devices, but they handle emotionally charged, adversarial bargaining poorly. EmoMAS is a framework that treats emotional expression as a strategic choice rather than a reaction. A Bayesian Orchestrator combines the emotion recommendations of three specialist agents:

  • a game-theoretic agent using Win-Stay, Lose-Shift,
  • a tabular Q-learning agent,
  • a psychological-coherence agent.

The orchestrator updates each agent's reliability online during the dialogue, so no pre-training is needed. The authors also release four synthetic, privacy-safe benchmarks covering debt collection, surgical scheduling, a child-companion robot and disaster rescue. In agent-to-agent simulations, both SLMs and LLMs equipped with EmoMAS outperformed all baselines on negotiation outcomes while keeping behavior ethically balanced. The work suggests that strategic emotion management can partly close the SLM–LLM gap. That makes private, on-device negotiation more feasible for high-stakes settings, although it has not yet been tested with real humans.

Key points

  • Emotion recognition and state space: Emotions are recognized via in-context prompting over seven discrete states (joy, sadness, anger, fear, disgust, surprise, neutral). The orchestrator fuses agent predictions with a softmax at temperature τ = 0.1.
  • Dual-signal reward: Final negotiation outcomes update agent reliabilities across dialogues. Within-turn agreement between agents adjusts weights inside a single dialogue. Together, these address the problem of sparse, delayed rewards.
  • Benchmarks: CRAD and SSD cover institution-local finance and medical negotiation. SSAD and DESRD cover embodied home-companion and disaster-rescue robots. All dialogues are LLM-generated (gpt-4o-mini).
  • Definition of SLM: Models with at most 7B parameters.
  • Limitations: Experiments are English-only and fully simulated, and the emotion set is fixed and discrete. The authors call for human oversight and safety filters before any real deployment. Code is available at github.com/Yunbo-max/EmoMAS.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Fields, topics and concepts

📚 Shelves 4The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksPsychology · 1,552 worksExperimental and Cognitive Psychology · 870 works

🧭 Topics 1Research subjects and languages the work is about.Sentiment & emotion · 972 works

Cite

Cite as Yunbo Long, Yuhan Liu and Liming Xu. 2026. EmoMAS: Emotion-Aware Multi-Agent System for High-Stakes Edge-Deployable Negotiation with Bayesian Orchestration. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers).

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.acl-long.1464

@inproceedings{long2026emomas,
  title = {{EmoMAS: Emotion-Aware Multi-Agent System for High-Stakes Edge-Deployable Negotiation with Bayesian Orchestration}},
  author = {Yunbo Long, Yuhan Liu and Liming Xu},
  booktitle = {Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},
  year = {2026},
  doi = {10.18653/v1/2026.acl-long.1464},
  url = {https://doi.org/10.18653/v1/2026.acl-long.1464}
}

References

31 references, 6 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Sahar Abdelnabi, Amr Gomaa, Sarath Sivaprasad, Lea Schönherr, and Mario Fritz. 2023. Llm-deliberation: Evaluating llms with interactive multi-agent negotiation games. arXiv preprint arXiv:2309.17234. — [arXiv:2309.17234](https://arxiv.org/abs/2309.17234) find
  2. Peter Belcak, Greg Heinrich, Shizhe Diao, Yonggan Fu, Xin Dong, Saurav Muralidharan, Yingyan Celine Lin, and Pavlo Molchanov. 2025. Small language models are the future of agentic ai. arXiv preprint arXiv:2506.02153. find
  3. Zhiwei Chen, Yupeng Hu, Zhiheng Fu, Zixu Li, Jiale Huang, Qinlei Huang, and Yinwei Wei. 2026. Intent: Invariance and discrimination-aware noise mitigation for robust composed image retrieval. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 40, pages 20463–20471. find
  4. Zhiwei Chen, Yupeng Hu, Zixu Li, Zhiheng Fu, Xuemeng Song, and Liqiang Nie. 2025. Offset: Segmentation-based focus shift revision for composed image retrieval. In Proceedings of the ACM International Conference on Multimedia, pages 6113– 6122. find
  5. Yunbo Long, Yuhan Liu, Liming Xu, and Alexandra Brintrup. 2026. EmoDebt: Bayesian-optimized emotional intelligence for strategic agent-to-agent debt recovery. In Proceedings of the 25th International Conference on Autonomous Agents and Multi-Agent Systems (AAMAS). find
  6. Zihao Cheng, Zeming Liu, Yingyu Shan, Xinyi Wang, Xiangrong Zhu, Yunpu Ma, Hongru Wang, Yuhang Guo, Wei Lin, and Yunhong Wang. 2026. Mem 2 evolve: Towards self-evolving agents via co-evolutionary capability expansion and experience distillation. Preprint, arXiv:2604.10923.
  7. Yunbo Long, Liming Xu, Lukas Beckenbauer, Yuhan Liu, and Alexandra Brintrup. 2025b. EvoEmo: Towards evolved emotional policies for adversarial llm agents in multi-turn price negotiation. arXiv preprint arXiv:2509.04310. — [arXiv:2509.04310](https://arxiv.org/abs/2509.04310) find
  8. Gerard Debreu. 1952. A social equilibrium existence theorem. Proceedings of the national academy of sciences, 38(10):886–893. find
  9. Zhenyan Lu, Xiang Li, Dongqi Cai, Rongjie Yi, Fangming Liu, Xiwen Zhang, Nicholas D Lane, and Mengwei Xu. 2024. Small language models: Survey, measurements, and insights. arXiv preprint arXiv:2409.15790.
  10. Zhezheng Hao, Hong Wang, Jian Luo, Jianqing Zhang, Yuyan Zhou, Qiang Lin, Can Wang, Hande Dong, and Jiawei Chen. 2026. Recreate: Reasoning and creating domain agents driven by experience. arXiv preprint arXiv:2601.11100.
  11. Feng He, Tianqing Zhu, Dayong Ye, Bo Liu, Wanlei Zhou, and Philip Yu. 2024. The emerged security and privacy of llm agent: A survey with case studies. ACM Computing Surveys. find
  12. Linmei Hu, Hongyu He, Duokang Wang, Ziwang Zhao, Yingxia Shao, and Liqiang Nie. 2024. Llm vs small model? large language model based text augmentation enhanced personality detection model. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, pages 18234–18242. find
  13. Yupeng Hu, Zixu Li, Zhiwei Chen, Qinlei Huang, Zhiheng Fu, Mingzhu Xu, and Liqiang Nie. 2026. Refine: Composed video retrieval via shared and differential semantics enhancement. ACM Transactions on Multimedia Computing, Communications and Applications. find
  14. Mengdi Li, Jiaye Lin, Xufeng Zhao, Wenhao Lu, Peilin Zhao, Stefan Wermter, and Di Wang. 2025. Curriculum-rlaif: Curriculum alignment with reinforcement learning from ai feedback. arXiv preprint arXiv:2505.20075. find
  15. Jonathan Light, Min Cai, Sheng Shen, and Ziniu Hu. 2023. AvalonBench: Evaluating llms playing the game of avalon. arXiv preprint arXiv:2310.05036. — [arXiv:2310.05036](https://arxiv.org/abs/2310.05036) find
  16. Jiaye Lin, Yifu Guo, Yuzhen Han, Sen Hu, Ziyi Ni, Licheng Wang, Mingguang Chen, Hongzhang Liu, Ronghao Chen, Yangfan He, and 1 others. 2025. Seagent: Self-evolution trajectory optimization in multi-step reasoning with llm-based agents. arXiv preprint arXiv:2508.02085. — [arXiv:2508.02085](https://arxiv.org/abs/2508.02085) find
  17. Yunbo Long, Yuhan Liu, and Alexandra Brintrup. 2025a. Eq-negotiator: Dynamic emotional personas empower small language models for edge-deployable credit negotiation. arXiv preprint arXiv:2511.03370. find
  18. Shichao Ma, Yunhe Guo, Jiahao Su, Qihe Huang, Zhengyang Zhou, and Yang Wang. 2026. Talk2image: A multi-agent system for multi-turn image generation and editing. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 40, pages 32437–32445. find
  19. Zeynep Örpek, Büşra Tural, and Zeynep Destan. 2024. The language model revolution: Llm and slm analysis. In 2024 8th International Artificial Intelligence and Data Processing Symposium (IDAP), pages 1–4. IEEE. find
  20. Jesse Prinz. 2004. Which emotions are basic. Emotion, evolution, and rationality, 69:88. find
  21. Johannes Schneider, Steffi Haag, and Leona Chandra Kruse. 2024. Negotiating with llms: Prompt hacks, skill gaps, and reasoning deficits. In International Conference on Computer-Human Interaction Research and Applications, pages 238–259. Springer. find
  22. Rui Sun, Jie Ding, Chenghua Gong, Tianjun Gu, Yihang Jiang, Juyuan Zhang, Liming Pan, and Linyuan Lü. 2026. Topodim: One-shot topology generation of diverse interaction modes for multi-agent systems. Preprint, arXiv:2601.10120.
  23. Junlin Wang, Jue Wang, Ben Athiwaratkun, Ce Zhang, and James Zou. 2024. Mixture-of-agents enhances large language model capabilities. arXiv preprint arXiv:2406.04692. find
  24. Yibo Yan, Shen Wang, Jiahao Huo, Philip S Yu, Xuming Hu, and Qingsong Wen. 2025. MathAgent: Leveraging a mixture-of-math-agent framework for real-world multimodal mathematical error detection. arXiv preprint arXiv:2503.18132. find
  25. Mingyu Zhang, Zixu Li, Zhiwei Chen, Zhiheng Fu, Xiaowei Zhu, Jiajia Nie, Yinwei Wei, and Yupeng Hu. 2026a. Hint: Composed image retrieval with dual-path compositional contextualized network. arXiv preprint arXiv:2603.26341. — [arXiv:2603.26341](https://arxiv.org/abs/2603.26341) find
  26. Peixuan Zhang, Chang Zhou, Ziyuan Zhang, Hualuo Liu, Chunjie Zhang, Jingqi Liu, Xiaohui Zhou, Xi Chen, Shuchen Weng, Si Li, and Boxin Shi. 2026b. A benchmark and multi-agent system for instruction-driven cinematic video compilation. Preprint, arXiv:2604.10456. — [arXiv:2604.10456](https://arxiv.org/abs/2604.10456) find
  27. Wenyuan Zhang, Xinghua Zhang, Haiyang Yu, Shuaiyi Nie, Bingli Wu, Juwei Yue, Tingwen Liu, and Yongbin Li. 2026c. Expseek: Self-triggered experience seeking for web agents. Preprint, arXiv:2601.08605.
  28. Xinglang Zhang, Yunyao Zhang, ZeLiang Chen, Junqing Yu, Wei Yang, and Zikai Song. 2026d. Logical phase transitions: Understanding collapse in llm logical reasoning. Preprint, arXiv:2601.02902.
  29. Yuanjun Zhang, Fuzel Ahamed Shaik, Suvojit Acharjee, Fahad Khalid, and Mourad Oussalah. 2026e. Towards reliable multimodal disaster severity assessment through preference optimization and explainable vision-language reasoning. Reliability Engineering & System Safety, page 112674. find
  30. Yunyao Zhang, Yihao Ai, Zuocheng Ying, Qirui Mi, Junqing Yu, Wei Yang, and Zikai Song. 2026f. Coupling macro dynamics and micro states for long-horizon social simulation. Preprint, arXiv:2604.05516. — [arXiv:2604.05516](https://arxiv.org/abs/2604.05516) find
  31. Shenzhe Zhu, Jiao Sun, Yi Nian, Tobin South, Alex Pentland, and Jiaxin Pei. 2025. The automated but risky game: Modeling agent-to-agent negotiations and transactions in consumer markets. In ICML 2025 Workshop on Reliable and Responsible Foundation Models. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/long-2026-emomas-emotion-aware-multi-agent/