Elamur Research⌕
☰
home / papers / 2026 / EmoRes: Toward Adaptive Psychological Support via User-Agnostic Benchmark and Topic-Mining Agent

EmoRes: Toward Adaptive Psychological Support via User-Agnostic Benchmark and Topic-Mining Agent

Paper · 2026 · Zhengwei Zou, Xuanming Jiang, Baoyi An et al.

Summary

LLM-based counseling chatbots tend to drift over long conversations. They repeat stock phrases, miss emotional signals from earlier turns, and give polite but shallow empathy. Evaluation is also difficult because most existing benchmarks are small or built on sensitive real patient data. The authors address both problems.

First, they build VeilEval, a bilingual benchmark that needs no real user data. It is produced in three stages. Variables are extracted from standardized clinical questionnaires. GPT-4o then generates synthetic client profiles and Q&A from those variables. Finally, the Q&A pairs are filtered by an LLM and by human experts.

Second, they propose EmoRes, a multi-agent framework based on Solution-Focused Brief Therapy. A topic-mining agent tracks weighted conversational themes from turn to turn. A self-reflection agent iteratively revises draft replies from both the counselor's and the client's viewpoint.

Across open-source, commercial and fine-tuned backbones, EmoRes yields gains of up to roughly three times over baselines without retraining. These gains are supported by ablations and human judgments. The work offers a reproducible, privacy-safe route to more coherent psychological-support dialogue.

Key points

  • VeilEval profiles are grounded in eight clinical scales: PHQ-9, GAD-7, PANSS, BSL-23, MDQ, PCL-5, LSAS and ASRS. Together they cover depression, anxiety, schizophrenia, borderline disorder, bipolar disorder, PTSD, social anxiety and ADHD.
  • Benchmark construction has three steps:
    • semi-synthetic data collection from the scales;
    • profile and contextual Q&A generation, covering background, psychological status and treatment goals;
    • multi-level filtering for generality, clinical relevance and contextual consistency.
  • EmoRes pairs two agents in one-to-one coordination:
    • TEAgent outputs topic labels with importance weights at every turn;
    • the Self-Reflection Agent refines each response, up to a fixed number of iterations.
  • Results: EmoRes improves on strong baselines by up to about 3× on VeilEval. Removing either agent hurts performance.
  • Limitations: GPT-4o-mini serves as both the simulated user and the evaluator, which may bias the results. Multi-agent coordination also adds response latency.

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

 Institutions

🎓Xi'an Jiaotong UniversityXi'an · China42 works · 36,431 citations🎓Shenzhen MSU-BIT UniversityShenzhen · China5 works · 38 citations

 Fields, topics and concepts

📚 Shelves 7The field of science this work belongs to (OpenAlex fields).Computer Science · 18,614 worksArtificial Intelligence · 17,965 worksArts and Humanities · 17,450 worksLanguage and Linguistics · 17,296 worksPsychology · 1,649 worksExperimental and Cognitive Psychology · 930 worksApplied Psychology · 25 works

🧭 Topics 1Research subjects and languages the work is about.Evaluation & benchmarks · 1,341 works

Cite

Cite as Zhengwei Zou, Xuanming Jiang, Baoyi An, Dingyu Nie, Zhengxing Fang, Qingyu Liu, Xueming Qian, Guoshuai Zhao and Zhongyu Yang. 2026. EmoRes: Toward Adaptive Psychological Support via User-Agnostic Benchmark and Topic-Mining Agent. Findings of the Association for Computational Linguistics: ACL 2026.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.findings-acl.416

@inproceedings{zou2026emores,
  title = {{EmoRes: Toward Adaptive Psychological Support via User-Agnostic Benchmark and Topic-Mining Agent}},
  author = {Zhengwei Zou, Xuanming Jiang, Baoyi An, Dingyu Nie, Zhengxing Fang, Qingyu Liu, Xueming Qian, Guoshuai Zhao and Zhongyu Yang},
  booktitle = {Findings of the Association for Computational Linguistics: ACL 2026},
  year = {2026},
  doi = {10.18653/v1/2026.findings-acl.416},
  url = {https://doi.org/10.18653/v1/2026.findings-acl.416}
}

References

56 references, 10 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Anthropic. 2024. Claude 3 haiku: our fastest model yet. https://www.anthropic.com/news/ claude-3-haiku. Accessed: 2025-10-07. ↗
  2. Fredrike P Bannink. 2007. Solution-focused brief therapy. Journal of Contemporary Psychotherapy, 37(2):87–94. find
  3. Clare Beatty, Tanya Malik, Saha Meheli, and Chaitali Sinha. 2022. Evaluating the therapeutic alliance with a free-text cbt conversational agent (wysa): A mixed-methods study. Frontiers in Digital Health, 4:847991. find
  4. Arun Brahma. 2024. Finetuning of falcon-7b llm using qlora on mental health conversational dataset. find
  5. Yekta Said Can and Cem Ersoy. 2023. Smart Affect Monitoring With Wearables in the Wild: An Unobtrusive Mood-Aware Emotion Recognition System . IEEE Transactions on Affective Computing, 14(04):2851–2863. find
  6. Rita Charon. 2008. Narrative medicine: Honoring the stories of illness. Oxford University Press. find
  7. Kai Chen and Zebing Sun. 2025. Deeppsy-agent: A stage-aware and deep-thinking emotional support agent system. Preprint, arXiv:2503.15876. — [arXiv:2503.15876](https://arxiv.org/abs/2503.15876) find
  8. Yujia Chen, Changsong Li, Yiming Wang, Tianjie Ju, Qingqing Xiao, Nan Zhang, Zifan Kong, Peng Wang, and Binyu Yan. 2025. MIND: Towards immersive psychological healing with multi-agent inner dialogue. In Findings of the Association for Computational Linguistics: EMNLP 2025, pages 9380–9413, Suzhou, C… find
  9. Zhiyu Chen, Yujie Lu, and William Wang. 2023. Empowering psychotherapy with large language models: Cognitive distortion detection through diagnosis of thought prompting. In Findings of the Association for Computational Linguistics: EMNLP 2023, pages 4295–4304, Singapore. Association for Computationa… find
  10. KuanChao Chu, Yi-Pei Chen, and Hideki Nakayama. 2024. Cohesive conversations: Enhancing authenticity in multi-agent simulated dialogues. In Proceedings of the 1st Conference on Language Modeling (COLM), Philadelphia, PA, USA. OpenReview. find
  11. Google Cloud. 2025. Gemini 2.5 flash-lite. https://cloud.google.com/vertex-ai/generative-ai/ Ac-docs/models/gemini/2-5-flash-lite. cessed: 2025-10-07. ↗
  12. Pierre Colombo, Wojciech Witon, Ashutosh Modi, James Kennedy, and Mubbasir Kapadia. 2019. Affect-driven dialog generation. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Paper… find
  13. Nicholas Z. Counts, David E. Bloom, and Neal Halfon. 2023. Psychological distress as a systemic economic risk in the usa. Nature Mental Health, 1:950–955. find
  14. Wenhui Cui, Christopher Michael Sandino, Hadi Pouransari, Ran Liu, Juri Minxha, Ellen L Zippi, Erdrin Azemi, and Behrooz Mahasseni. 2026. Em-bridge: Enhancing gesture generalization from emg signals through cross-modal representation learning. In The Fourteenth International Conference on Learning R… find
  15. DeepSeek-AI. 2024. Deepseek-v3 technical report. Preprint, arXiv:2412.19437. find
  16. DeepSeek-AI. 2025. Deepseek-r1: Incentivizing reasoning capability in llms via reinforcement learning. Preprint, arXiv:2501.12948.
  17. Stéphane d’Ascoli, Corentin Bel, Jérémy Rapin, Hubert Banville, Yohann Benchetrit, Christophe Pallier, and Jean-Rémi King. 2025. Towards decoding individual words from non-invasive brain recordings. Nature Communications, 16(1):10521. find
  18. Musthafa Mohamed Firose, Bogahawaththage Nishadi Madushika Chathurangi, and Imriyas Kamardeen. 2025. Work stress among construction professionals during an economic crisis: a case study of sri lanka. Smart and Sustainable Built Environment. find
  19. C. Franklin, K. W. Bolton, and S. Guz. 2019. Solution-focused brief family therapy. In B. H. Fiese, M. Celano, K. Deater-Deckard, E. N. Jouriles, and M. A. Whisman, editors, APA handbook of contemporary family psychology: Family therapy and training, pages 139–153. American Psychological Association… find
  20. C. Franklin, A. Zhang, A. S. Froerer, and S. K. Johnson. 2017. Solution focused brief therapy: A systematic review and meta-summary of process research. Journal of Marital and Family Therapy, 43(1):16–30. find
  21. Cynthia Franklin, Terry S. Trepper, Eric E. McCollum, and Wallace J. Gingerich. 2011. Solution-Focused Brief Therapy: A Handbook of Evidence-Based Practice. Oxford University Press. find
  22. Muskan Garg, Amirmohammad Shahbandegan, Amrit Chadha, and Vijay Mago. 2023. An annotated dataset for explainable interpersonal risk factors of mental disturbance in social media posts. In Findings of the Association for Computational Linguistics: ACL 2023, pages 11960–11969, Toronto, Canada. Associa… find
  23. Team GLM, Aohan Zeng, Bin Xu, Bowen Wang, Chenhui Zhang, Da Yin, Diego Rojas, Guanyu Feng, Hanlin Zhao, Hanyu Lai, Hao Yu, Hongning Wang, Jiadai Sun, Jiajie Zhang, Jiale Cheng, Jiayi Gui, Jie Tang, Jing Zhang, Juanzi Li, and 37 others. 2024. Chatglm: A family of large language models from glm-130b t…
  24. Wei Gu, Luo Tianming, Qiran Zhang, Mohan Ye, Xiao Shen, Wenxin Chen, Yunhuan Li, Yichen Zhang, Jing Hong, Bao-liang Lu, and 1 others. 2026. Cere-bragloss: Instruction-tuning a large vision-language model for fine-grained clinical eeg interpretation. In The Fourteenth International Conference on Lear… find
  25. Amey Hengle, Atharva Kulkarni, Shantanu Deepak Patankar, Madhumitha Chandrasekaran, Sneha D’silva, Jemima S. Jacob, and Rashmi Gupta. 2024. Still not quite there! evaluating large language models for comorbid mental health diagnosis. In Proceedings of the 2024 Conference on Empirical Methods in Natu… find
  26. Mengkang Hu, Tianxing Chen, Qiguang Chen, Yao Mu, Wenqi Shao, and Ping Luo. 2025. HiAgent: Hierarchical working memory management for solving long-horizon agent tasks with large language model. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long…
  27. Songbo Hu, Han Zhou, Mete Hergul, Milan Gritta, Guchun Zhang, Ignacio Iacobacci, Ivan Vulić, and Anna Korhonen. 2023. Multi3woz: A multilingual, multi-domain, multi-parallel dataset for training and evaluating culturally adapted task-oriented dialog systems. Transactions of the Association for Comp… find
  28. Wei-Bang Jiang, Yansen Wang, Bao-Liang Lu, and Dongsheng Li. 2025a. NeuroLM: A universal multi-task foundation model for bridging the gap between language and EEG signals. In The Thirteenth International Conference on Learning Representations, Singapore. OpenReview. find
  29. Xuanming Jiang, Baoyi An, Zhengwei Zou, Dingyu Nie, Jialie Shen, Xueming Qian, and Guoshuai Zhao. 2025b. Ear with eye: Lightweight multimodal audiovisual network inspired by bionic structures. In Proceedings of the 33rd ACM International Conference on Multimedia, page 1346–1355, New York, NY, USA. A… find
  30. Xuanming Jiang, Dingyu Nie, Baoyi An, Yuzhe Zheng, Yichuan Mao, Jialie Shen, Xueming Qian, Zhiwen Jin, Wei Lan, and Guoshuai Zhao. 2026. Whole-field action sensing via wearable single-channel emg sensors and resource-efficient motion network. Proceedings of the AAAI Conference on Artificial Intellig… find
  31. Tin Lai, Yukun Shi, Zicong Du, Jiajie Wu, Ken Fu, Yichao Dou, and Ziqi Wang. 2023. Psy-llm: Scaling up global mental health psychological services with ai-based large language models. Preprint, arXiv:2307.11991. — [arXiv:2307.11991](https://arxiv.org/abs/2307.11991) find
  32. Kunyao Lan, Bingrui Jin, Zichen Zhu, Siyuan Chen, Shu Zhang, Kenny Q Zhu, and Mengyue Wu. 2024. Depression diagnosis dialogue simulation: Self-improving psychiatrist with tertiary memory. Preprint, arXiv:2409.15084. — [arXiv:2409.15084](https://arxiv.org/abs/2409.15084) find
  33. Suyeon Lee, Sunghwan Kim, Minju Kim, Dongjin Kang, Dongil Yang, Harim Kim, Minseok Kang, Dayi Jung, Min Hee Kim, Seungbeen Lee, Kyong-Mee Chung, Youngjae Yu, Dongha Lee, and Jinyoung Yeo. 2024a. Cactus: Towards psychological counseling conversations using cognitive behavioral theory. In Findings of… find
  34. Yeonji Lee, Sangjun Park, Kyunghyun Cho, and JinYeong Bak. 2024b. Mentalagora: A gateway to advanced personalized care in mental health through multi-agent debating and attribute control. Preprint, arXiv:2407.02736. find
  35. Guohao Li, Hasan Hammoud, Hani Itani, Dmitrii Khizbullin, and Bernard Ghanem. 2023. Camel: Communicative agents for "mind" exploration of large language model society. In Advances in Neural Information Processing Systems, pages 51991– 52008, Red Hook, NY, USA. Curran Associates, Inc.
  36. Ruoqi Liu, Yuelin Bai, Xiang Yue, and Ping Zhang. 2026a. Teaching multimodal llms to comprehend 12-lead electrocardiographic images. npj Digital Medicine. find
  37. Yunsong Liu, Zunamys I Carrero, Xiaofeng Jiang, Dyke Ferber, Georg Wölflein, Li Zhang, Sanddhya Jaya-balan, Tim Lenz, Zhouguang Hui, and Jakob Nikolas Kather. 2026b. Benchmarking large language model-based agent systems for clinical decision tasks. npj Digital Medicine. find
  38. Zengding Liu, Chen Chen, Jiannong Cao, Minglei Pan, Jikui Liu, Nan Li, Fen Miao, and Ye Li. 2024. Large language models for cuffless blood pressure measurement from wearable biosignals. In Proceedings of the 15th ACM International Conference on Bioinformatics, Computational Biology and Health Inform… find
  39. Jingping Nie, Hanya (Vera) Shao, Yuang Fan, Qijia Shao, Haoxuan You, Matthias Preindl, and Xiaofan Jiang. 2025. Llm-based conversational ai therapist for daily functioning screening and psychotherapeutic intervention via everyday smart devices. ACM Transactions on Computing for Healthcare. find
  40. OpenAI. 2022. Chatgpt: Optimizing language models for dialogue. Accessed: 2025-03-04. find
  41. Huachuan Qiu, Hongliang He, Shuai Zhang, Anqi Li, and Zhenzhong Lan. 2024. SMILE: Single-turn to multi-turn inclusive language expansion via ChatGPT for mental health support. In Findings of the Association for Computational Linguistics: EMNLP 2024, pages 615–636, Miami, Florida, USA. Association fo…
  42. Huachuan Qiu and Zhenzhong Lan. 2024. Interactive agents: Simulating counselor-client psychological counseling via role-playing llm-to-llm interactions. Preprint, arXiv:2408.15787.
  43. Nishat Raihan, Sadiya Sayara Chowdhury Puspo, Shafkat Farabi, Ana-Maria Bucur, Tharindu Ranasinghe, and Marcos Zampieri. 2024. MentalHelp: A multi-task dataset for mental health in social media. In Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resource… find
  44. EmoLLM Team. 2024. Emollm: Reinventing mental health support with large language models. find
  45. Hugo Touvron and 1 others. 2024. The llama 3 herd of models. Preprint, arXiv:2407.21783.
  46. Sichang Tu, Abigail Powers, Natalie Merrill, Negar Fani, Sierra Carter, Stephen Doogan, and Jinho D. Choi. 2024. Automating PTSD diagnostics in clinical interviews: Leveraging large language models for trauma assessments. In Proceedings of the 25th Annual Meeting of the Special Interest Group on Dis… find
  47. Shenghan Wu, Yimo Zhu, Wynne Hsu, Mong-Li Lee, and Yang Deng. 2025. From personas to talks: Revisiting the impact of personas on LLM-synthesized emotional support conversations. In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, pages 5439–5453, Suzhou, China.… find
  48. Mengxi Xiao, Qianqian Xie, Ziyan Kuang, Zhicheng Liu, Kailai Yang, Min Peng, Weiguang Han, and Jimin Huang. 2024. HealMe: Harnessing cognitive reframing in large language models for psychotherapy. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: L… find
  49. Ancheng Xu, Di Yang, Renhao Li, Jingwei Zhu, Minghuan Tan, Min Yang, Wanxin Qiu, Mingchen Ma, Haihong Wu, Bingyu Li, and 1 others. 2025. Autocbt: An autonomous multi-agent framework for cognitive behavioral therapy in psychological counseling. Preprint, arXiv:2501.09426. — [arXiv:2501.09426](https://arxiv.org/abs/2501.09426) find
  50. An Yang, Anfeng Li, Baosong Yang, Beichen Zhang, Binyuan Hui, Bo Zheng, Bowen Yu, Chang Gao, Chengen Huang, Chenxu Lv, Chujie Zheng, Dayiheng Liu, Fan Zhou, Fei Huang, Feng Hu, Hao Ge, Haoran Wei, Huan Lin, Jialong Tang, and 41 others. 2025a. Qwen3 technical report. Preprint, arXiv:2505.09388.
  51. Zhongyu Yang, Wei Pang, and Yingfang Yuan. 2026a. Xr: Cross-modal agents for composed image retrieval. In Proceedings of the ACM Web Conference 2026, WWW ’26, page 2071–2082, New York, NY, USA. Association for Computing Machinery. find
  52. Zhongyu Yang, Junhao Song, Siyang Song, Wei Pang, and Yingfang Yuan. 2025b. MERMAID: Multi-perspective self-reflective agents with generative augmentation for emotion recognition. In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, pages 24650–24666, Suzhou, Ch… find
  53. Zhongyu Yang, Zuhao Yang, Shuo Zhan, Tan Yue, Wei Pang, and Yingfang Yuan. 2026b. Svagent: Storyline-guided long video understanding via cross-modal multi-agent collaboration. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Denver CO, United States. IEEE. find
  54. Zhongyu Yang, Yingfang Yuan, Xuanming Jiang, Baoyi An, and Wei Pang. 2026c. Inex: Hallucination mitigation via introspection and cross-modal multi-agent collaboration. Proceedings of the AAAI Conference on Artificial Intelligence, 40(35):29829–29837. find
  55. Mian Zhang, Shaun M Eack, and Zhiyu Zoey Chen. 2025. Preference learning unlocks llms’ psycho-counseling skills. Preprint, arXiv:2502.19731.
  56. Jinfeng Zhou, Yuxuan Chen, Jianing Yin, Yongkang Huang, Yihan Shi, Xikun Zhang, Libiao Peng, Rongsheng Zhang, Tangjie Lv, Zhipeng Hu, Hongning Wang, and Minlie Huang. 2025. Crisp: Cognitive restructuring of negative thoughts through multi-turn supportive dialogues. In Proceedings of the 2025 Confere…

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/zou-2026-emores-adaptive-psychological-support-user/