AI-BASED PERSONALIZED LEARNING PLATFORM

AI-based personalized learning has long suffered from a tension between the engineering elegance of its components and the pedagogical effectiveness of their integration. Knowledge tracing, content recommendation, and generative tutoring have each produced strong individual results yet have rarely been combined in ways that explicitly model the cognitive state of the learner moment to moment. ADAPT-CL proposes one such integration, anchored on the observation that interaction telemetry alone, suitably engineered, supplies a workable signal of cognitive load without specialized hardware. A pilot study supports the framework's viability and suggests effect sizes large enough to motivate substantial replication. We do not claim that ADAPT-CL is the final word on personalized learning—our limitations are numerous and our population narrow—but we believe the design principles it articulates, namely that personalization is a joint optimization problem over mastery, cognitive load, and generative content, will outlast the particular implementation choices reported here. The two-sigma problem has resisted technological solution for forty years; it will not be solved by any single component, however impressive. It will be solved, if at all, by the careful orchestration of many.  

1.       Aleven, V., McLaughlin, E. A., Glenn, R. A., & Koedinger, K. R. (2016). Instruction based on adaptive learning technologies. In R. E. Mayer & P. Alexander (Eds.), Handbook of research on learning and instruction (2nd ed., pp. 522–560). Routledge. 2.       Anderson, J. R., Corbett, A. T., Koedinger, K. R., & Pelletier, R. (1995). Cognitive tutors: Lessons learned. The Journal of the Learning Sciences, 4(2), 167–207. 3.       Baker, R. S. J. d., D'Mello, S. K., Rodrigo, M. M. T., & Graesser, A. C. (2008). Better to be frustrated than bored: The incidence, persistence, and impact of learners' cognitive–affective states during interactions with three different computer-based learning environments. International Journal of Human-Computer Studies, 68(4), 223–241. 4.       Bastani, H., Bastani, O., Sungu, A., Ge, H., Kabakcı, Ö., & Mariman, R. (2024). Generative AI can harm learning. SSRN Working Paper 4895486. 5.       Bloom, B. S. (1984). The 2 sigma problem: The search for methods of group instruction as effective as one-to-one tutoring. Educational Researcher, 13(6), 4–16. 6.       Carbonell, J. R. (1970). AI in CAI: An artificial-intelligence approach to computer-assisted instruction. IEEE Transactions on Man-Machine Systems, 11(4), 190–202. 7.       Clement, B., Roy, D., Oudeyer, P. Y., & Lopes, M. (2015). Multi-armed bandits for intelligent tutoring systems. Journal of Educational Data Mining, 7(2), 20–48. 8.       Corbett, A. T., & Anderson, J. R. (1995). Knowledge tracing: Modeling the acquisition of procedural knowledge. User Modeling and User-Adapted Interaction, 4(4), 253–278. 9.       Ghosh, A., Heffernan, N., & Lan, A. S. (2020). Context-aware attentive knowledge tracing. Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, 2330–2339. 10.    Hart, S. G., & Staveland, L. E. (1988). Development of NASA-TLX (Task Load Index): Results of empirical and theoretical research. Advances in Psychology, 52, 139–183.