Embedded Files
Challenge
Challenge
Supporting middle-school mathematics practice in class and at home.
Supporting middle-school mathematics practice in class and at home.
Students need feedback and support while solving mathematics problems independently.
Mobile access offers another way to practice outside school.
Parents and caregivers need practical ways to support homework without taking over.
What We Do
What We Do
“Optimizing AI-based tutoring software for middle-school mathematics on smartphones” studies how tutoring software, mobile access, and support from teachers and caregivers shape independent mathematics practice.
The project combines classroom experiments with design studies of learner control, goal support, and caregiver involvement. Data-driven tutor revisions showed limited benefits in process measures in Experiment 2, without corresponding improvements on pre/post tests.
Vincent Aleven leads the project with co-PIs Ken Koedinger and Paulo Carvalho, and collaborators including Jonathan Sewall, Qianru Lyu, and Conrad Borchers.
Supported by the Institute of Education Sciences, U.S. Department of Education, award R305A220386.
Educators interested in participating in CMU research: cats-lab@cmu.edu.
Key Findings
Key Findings
Tutoring Software and Paper Practice
Tutoring Software and Paper Practice
Students learned more and more efficiently with intelligent tutoring than with matched paper problems.
Students with lower initial math self-efficacy showed greater learning gains across both tutor and paper practice (Cen et al., 2026).
Caregiver Support
Caregiver Support
Caregivers appreciated AI-generated message suggestions that helped them prompt their child’s thinking and offer instructional support—even when they weren’t confident in the math content (Venugopalan et al., 2025).
Learning in Your Hands
Learning in Your Hands
Design studies examined how problem choice, skill-mastery displays, and “what-if” explanations shape students’ practice decisions.
“What-if” explanations shifted students’ focus toward skills nearing mastery, supporting interfaces highlighting achievable progress alongside practice choices (Borchers et al., 2025).
Goal-Driven Homework
Goal-Driven Homework
Goal support combines personalized targets, progress feedback, and teacher-supported rewards.
A Fall 2025 analysis associated goal support with 28% more weekly practice time and 52% more skills completed (Borchers, 2026).
Why This Matters
Why This Matters
Effective mathematics tutoring depends on more than feedback during problem solving. Students also need reasons to keep practicing and support beyond the classroom. This project brings these needs together through smartphone access, personalized practice goals, and tools for teachers and caregivers. The work advances AI tutoring toward a broader role: helping students learn from practice and helping the people around them sustain their effort.
Related Projects
Related Projects
Publications
Publications
Borchers, C., Carvalho, P. F., Xia, M., Liu, P., Koedinger, K. R., & Aleven, V. (2023). What makes problem-solving practice effective? Comparing paper and AI tutoring. In Viberg, I. Jivet, P. Muñoz-Merino, M. Perifanou, & T. Papathoma (Eds.), Proceedings 18th European Conference on Technology Enhanced Learning, EC-TEL 2023 (pp. 44-59). Cham, Switzerland: Springer. https://doi.org/10.1007/978-3-031-42682-7_4
Nguyen, H. T., Borchers, C., Xia, M., & Aleven, V. (2024). Designing tools for caregiver involvement in intelligent tutoring systems for middle school mathematics. Proceedings of the 2024 ISLS Annual Meeting. https://doi.org/10.22318/icls2024.630637
Peng, C., Borchers, C., & Aleven, V. (2024). Designing homework support tools for middle school mathematics using intelligent tutoring systems. Proceedings of the 2024 ISLS Annual Meeting. https://doi.org/10.22318/icls2024.989202
Venugopalan, D., Yan, Z., Borchers, C., Lin, J., & Aleven, V. (2025). Combining Large Language Models with Tutoring System Intelligence: A Case Study in Caregiver Homework Support. In LAK25: The 15th International Learning Analytics and Knowledge Conference (LAK 2025), March 03–07, 2025, Dublin, Ireland. ACM, New York, NY, USA. https://doi.org/10.1145/3706468.3706516
Borchers, C., Ooge, J., Peng, C., & Aleven, V. (2025). How Learner Control and Explainable Learning Analytics on Skill Mastery Shape Student Desires to Finish and Avoid Loss in Tutored Practice. In LAK25: The 15th International Learning Analytics and Knowledge Conference (LAK 2025), March 03–07, 2025, Dublin, Ireland. ACM, New York, NY, USA. https://doi.org/10.1145/3706468.3706484
Borchers, C., Nguyen, H. T., Carvalho, P. F., Koedinger, K. R., & Aleven, V. (2025, July). Involving Parents in Tutoring Systems to Increase Content Confidence: A Design Probe Study. In International Conference on Artificial Intelligence in Education (pp. 107-114). Cham: Springer Nature Switzerland.
Borchers, C., Peng, C., Lyu, Q., Carvalho, P. F., Koedinger, K. R., & Aleven, V. (2025). Student Perceptions of Adaptive Goal Setting Recommendations: A Design Prototyping Study. In Proceedings of the 26th International Conference on Artificial Intelligence in Education (AIED 2025). Palermo, Italy.
Borchers, C., & Shou, T. (2026). Can large language models match tutoring system adaptivity? A benchmarking study. Proceedings of the 26th International Conference on Artificial Intelligence in Education (AIED).
Borchers, C., Lyu, Q., Kanchi, R., Koedinger, K., Aleven, V., & Carvalho, P. Student Learning Rates When Practicing with an Intelligent Tutoring System Versus On Paper. OSF Preprints.
Borchers, C. (2026). Personalized Goal Support to Enhance Technology-Supported Learning (Doctoral dissertation). Carnegie Mellon University.
Page updated
Report abuse