CTAt

CTAT stands for Cognitive Tutor Authoring Tools, a software framework for building intelligent tutoring systems. It provides an integrated environment for designing, deploying, and evaluating cognitive tutors that adapt to individual learners across a range of domains.

The framework relies on constraint-based modeling and model tracing. Domain experts define a set of constraints that describe correct cognitive steps and a solution path. The tutor uses these constraints to assess student actions, provide immediate feedback, and guide learners with adaptive hints and scaffolding. Authors can model goals, knowledge components, and tutoring policies, while the system handles the interaction with students, data collection, and performance analysis.

CTAT originated in education research on intelligent tutoring systems, with development historically associated with Carnegie Mellon University and collaborators. It has been used in classroom studies and online learning to examine problem solving, procedural fluency, and conceptual understanding. The authoring tools enable educators and researchers to create domain-specific tutors without requiring extensive AI programming, and tutors can be deployed over the web and integrated with learning management systems.

Key features include an authoring environment for defining constraints, skills, and hints; model-tracing and constraint-based feedback; adaptive hints and stepwise scaffolding; diagnostic feedback for common student errors; data logging for analytics; and deployment support for web-based delivery. CTAT aims to make the construction of customizable, evidence-based tutors accessible to educators and researchers while supporting iterative design and evaluation.

The project provides documentation, tutorials, and example tutors to help users get started. CTAT-related resources cover domains such as mathematics and programming, and tutors created with CTAT can be run in standard web browsers, supporting research and instructional deployment.