LearnOpt: Recovering the Latent Cognitive Structure of Standardized Examinations via Knowledge Graphs and Constrained Optimization

LearnOpt is a novel framework that utilizes knowledge graphs and constrained optimization to recover the latent cognitive structure of standardized examinations, specifically applied to NEET questions over nine years (2016-2024). It constructs an exam knowledge graph from LLM-tagged questions, revealing a stable latent skill distribution that shifts significantly with changes in curriculum, as evidenced by KL divergence metrics. This work is significant for practitioners as it provides a method for generating personalized study plans based on latent cognitive structures, enhancing targeted learning strategies in educational contexts, and it includes public access to the code, knowledge graph, and annotated dataset.