ClinTutor-R1: Advancing Scalable and Robust One-to-Many Alignment in Clinical Socratic Education

Zhitao He, Haolin Yang, Zeyu Qin, Yi R. Fung
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:42146-42195, 2026.

Abstract

While Large Language Models (LLMs) have achieved remarkable success in dyadic (one-on-one) instruction, they face significant challenges in One-to-Many alignment, such as clinical ward rounds, where an instructor must simultaneously guide a diverse group of trainees. Current models often suffer from context dilution and goal misalignment, failing to balance individual scaffolding with collective learning progress. To address this, we introduce ClinEdu, a multi-agent pedagogical simulator that model the complexity of group dynamics. Leveraging this platform, we construct ClinTeach, a large-scale dataset of Socratic teaching dialogues, and propose ClinTutor-R1, the first multimodal agent explicitly architected to achieve one-to-many alignment in clinical education, employing an explicit internal thinking mechanism to model both individual belief states and group consensus. We validate our framework through a comprehensive protocol covering both standard static benchmarks and rigorous in-situ interactive evaluation within ClinEdu. Experimental results demonstrate that ClinTutor-R1 outperforms base models by over 20% and achieves parity with proprietary reasoning models , while exhibiting exceptional scalability in maintaining instructional quality across expanding student cohorts.

Cite this Paper

BibTeX
@InProceedings{pmlr-v306-he26as, title = {{C}lin{T}utor-R1: Advancing Scalable and Robust One-to-Many Alignment in Clinical Socratic Education}, author = {He, Zhitao and Yang, Haolin and Qin, Zeyu and Fung, Yi R.}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {42146--42195}, year = {2026}, editor = {Zhang, Tong and Dudik, Miroslav and Jaggi, Martin and Agarwal, Alekh and Li, Sharon and Schuurmans, Dale and Zhu, Jerry and Berkenkamp, Felix and Dong, Hanze and Bietti, Alberto}, volume = {306}, series = {Proceedings of Machine Learning Research}, month = {06--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v306/main/assets/he26as/he26as.pdf}, url = {https://proceedings.mlr.press/v306/he26as.html}, abstract = {While Large Language Models (LLMs) have achieved remarkable success in dyadic (one-on-one) instruction, they face significant challenges in One-to-Many alignment, such as clinical ward rounds, where an instructor must simultaneously guide a diverse group of trainees. Current models often suffer from context dilution and goal misalignment, failing to balance individual scaffolding with collective learning progress. To address this, we introduce ClinEdu, a multi-agent pedagogical simulator that model the complexity of group dynamics. Leveraging this platform, we construct ClinTeach, a large-scale dataset of Socratic teaching dialogues, and propose ClinTutor-R1, the first multimodal agent explicitly architected to achieve one-to-many alignment in clinical education, employing an explicit internal thinking mechanism to model both individual belief states and group consensus. We validate our framework through a comprehensive protocol covering both standard static benchmarks and rigorous in-situ interactive evaluation within ClinEdu. Experimental results demonstrate that ClinTutor-R1 outperforms base models by over 20% and achieves parity with proprietary reasoning models , while exhibiting exceptional scalability in maintaining instructional quality across expanding student cohorts.} }
Endnote
%0 Conference Paper %T ClinTutor-R1: Advancing Scalable and Robust One-to-Many Alignment in Clinical Socratic Education %A Zhitao He %A Haolin Yang %A Zeyu Qin %A Yi R. Fung %B Proceedings of the 43rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2026 %E Tong Zhang %E Miroslav Dudik %E Martin Jaggi %E Alekh Agarwal %E Sharon Li %E Dale Schuurmans %E Jerry Zhu %E Felix Berkenkamp %E Hanze Dong %E Alberto Bietti %F pmlr-v306-he26as %I PMLR %P 42146--42195 %U https://proceedings.mlr.press/v306/he26as.html %V 306 %X While Large Language Models (LLMs) have achieved remarkable success in dyadic (one-on-one) instruction, they face significant challenges in One-to-Many alignment, such as clinical ward rounds, where an instructor must simultaneously guide a diverse group of trainees. Current models often suffer from context dilution and goal misalignment, failing to balance individual scaffolding with collective learning progress. To address this, we introduce ClinEdu, a multi-agent pedagogical simulator that model the complexity of group dynamics. Leveraging this platform, we construct ClinTeach, a large-scale dataset of Socratic teaching dialogues, and propose ClinTutor-R1, the first multimodal agent explicitly architected to achieve one-to-many alignment in clinical education, employing an explicit internal thinking mechanism to model both individual belief states and group consensus. We validate our framework through a comprehensive protocol covering both standard static benchmarks and rigorous in-situ interactive evaluation within ClinEdu. Experimental results demonstrate that ClinTutor-R1 outperforms base models by over 20% and achieves parity with proprietary reasoning models , while exhibiting exceptional scalability in maintaining instructional quality across expanding student cohorts.
APA
He, Z., Yang, H., Qin, Z. & Fung, Y.R.. (2026). ClinTutor-R1: Advancing Scalable and Robust One-to-Many Alignment in Clinical Socratic Education. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:42146-42195 Available from https://proceedings.mlr.press/v306/he26as.html.

Related Material