intelligenter-tutoring-dialog

To design a tutoring dialogue that addresses misconceptions, you input a learning goal and anticipated difficulties to generate a branching dialog architecture with specific tutor moves and decision rules for resolving student errors.

The best way to structure branching intelligent tutoring dialogs is using a phased interaction model of eliciting, conflicting, and constructing reasoning, which helps navigate the assistance dilemma and maximize durable learning.

Yes, you can use this dialog design approach for one-on-one classroom scaffolding, as it produces a library of tutor moves and worked examples specifically applicable for classroom teachers and small-group secondary-level interactions.

You select the right tutor moves during a formative assessment conversation by applying explicit IF→THEN decision rules based on student responses, choosing from a catalog of questions, pumps, hints, silence, and minimal corrections.

To generate an intelligent tutoring system interaction policy, you need to provide inputs for the learning goal and anticipated difficulties, while optional context fields include student level, subject, interaction length, student model, and system capabilities.

You should not use a multi-turn tutoring dialog when a quick direct correction is sufficient, as this approach is designed for deeper conceptual misconceptions requiring phased conflict and construction to ensure durable learning rather than passive reception.