Jingzhe Lin, Hengbin Yu, Yongdan Zeng, Fangwei Zhong — ICML 2026 — cs.MA, cs.CY 📄 Full text (arXiv)
EduMirror introduces a multi-agent simulator for studying educational social dynamics, addressing the dilemma that observational studies lack causal power while controlled experiments face ethical constraints. The simulator features value-driven agents grounded in psychological needs and social value orientation, with a dual-track measurement protocol quantifying both observable behaviors and latent psychological states. Validation through case studies on school bullying and group cooperation, plus broader evaluations across diverse educational scenarios, shows the simulator generates realistic, theory-consistent dynamics measurable by empirical criteria. This enables structured in silico educational research for hypothesis testing and counterfactual intervention analysis. The approach connects to student-modeling by modeling latent psychological states and to learning-analytics by providing a sandbox for testing interventions before deployment in real classrooms.
Related Pages
- knowledge-tracing-irt — Knowledge tracing models and IRT for student modeling
- intelligent-tutoring — AI tutoring systems and adaptive instruction
- learning-analytics — Data-driven analysis of learning processes
- ai-literacy — Understanding and evaluating AI tools
- student-modeling — Representing learner knowledge and behavior
- cs-education — Computing education research and pedagogy
- self-regulated-learning — Metacognitive strategies for independent learning
- formative-assessment — Ongoing assessment to inform instruction
- automated-grading — AI-assisted evaluation of student work
- scaffolding — Instructional support that fades with competence
Citations
APA: Jingzhe Lin, Hengbin Yu, Yongdan Zeng, Fangwei Zhong (2026). EduMirror: Modeling Educational Social Dynamics with Value-driven Multi-agent Simulation. arXiv:2606.07948. ICML 2026 — cs.MA, cs.CY.