Capability profiles are precisely defined as structured cognitive states that systematically encode the specific skills, knowledge, and misconceptions (what students can and cannot yet demonstrate) associated with various performance levels in a given domain. In the context of the DrawSim-PD framework, they serve as a foundational mechanism to generate pedagogically imperfect, yet realistic, student-like artifacts. By explicitly defining these cognitive states, capability profiles ensure consistency and coherence across diverse output modalities—such as student drawings, reasoning narratives, and diagnostic concept maps—all reflecting a specific student's understanding. This approach is crucial for overcoming data scarcity and privacy concerns in educational research and professional development, enabling scalable teacher training by providing a rich, synthetic dataset for diagnostic reasoning practice. Researchers in AI in Education, educational psychology, and generative modeling utilize these profiles to create realistic learning simulations and assessment tools.
Capability profiles are structured descriptions of what students know and don't know at different skill levels. They are used in AI systems to create realistic, fake student work, like drawings or explanations, for teachers to practice diagnosing. This helps train teachers without needing to share real student data, which is often restricted by privacy rules.
student capability profiles, cognitive state profiles, performance profiles
Was this definition helpful?