Education Technology
Framework Uses Structured Ontologies to Build Better Geometry Lesson Plans
A decades-old approach to classifying geometry problems through interconnected facts and skills is being revived as educators and AI researchers seek new ways to teach logical thinking.
Geometry has long been prized for developing students' capacity for logical reasoning and abstract thought, yet its role in mathematics curricula has diminished over time. A resurgence of interest—partly fueled by progress in artificial intelligence and educational tools—is now drawing attention back to geometry's cognitive value and prompting innovation in how problems are solved and proofs are verified.
The Ontology-Based Framework#
Researchers have revived an ontology-based system, originally conceived in the 1990s, that organizes geometry problems, their solutions, and the skills they develop into a structured, interconnected knowledge map. Rather than treating problems as isolated exercises, this framework maps out facts, objects, and methods in a way that allows educators to understand which competencies each problem builds and how they relate to one another.
The framework's central innovation is the concept of "solution graphs": directed acyclic graphs that represent multiple pathways to solving a single problem and the chain of skills required at each step. This mapping enables instructors to align problem selection directly with what they want students to learn, rather than relying on intuition or generic textbook ordering.
Tracking Progress and Designing Curricula#
By encoding skill dependencies and solution strategies in this structured way, the system makes it possible to track student abilities at a granular level. Educators can see not just whether a student solved a problem, but which cognitive steps they mastered and where they faltered. Over three decades of practical use, the framework has been tested through the annotation of thousands of problems, demonstrating its viability at scale.
Challenges and Next Steps#
The researchers identify two major open problems that, if solved, would amplify the framework's utility: automated problem annotation and solution validation. Automating these tasks would significantly reduce the time teachers spend checking student work and would make it feasible to offer real-time, interactive feedback to self-directed learners. These advances could be particularly valuable as educators seek to reintegrate geometry into curricula and harness AI to personalize instruction.
