Skip to content

Intelligent cognitive tutor for planning in partially observable environments

Design · All designs
Evidence · 1 claim (1 for) · 1 study (1 causal), q3 · 0 of 1 report an effect size · 1 claim rests on one study

Description

An intelligent tutoring system that teaches MGPO-discovered planning strategies through interactive practice with metacognitive feedback. "To teach people the planning strategies discovered by MGPO, we developed a cognitive tutor" that presents learners with highlighted planning-action choices, provides feedback on suboptimal selections with delay penalties proportional to the VOC difference, and demonstrates the optimal strategy through step-by-step videos. Unlike prior tutors using dynamic programming with cached Q-values, this tutor computes VOC values online via MGPO, enabling application to large, partially observable environments where pre-computing feedback is infeasible.

Design Implications

Context

Requirements

  • The MGPO algorithm for online VOC computation
  • A set of precomputed observation samples for the environment

Constraints

  • Delay penalty uses a fixed 0.05 VOC threshold rather than exact Q-value differences
  • Premature termination triggers a delay penalty rather than actual termination when MGPO would continue planning

Target Learners

  • adult online learners

Learning Goals

  • planning strategies
  • metacognitive skill
  • resource-rational information search

Claims

Examples

-

Key Sources

  • Heindrich, L., Consul, S., & Lieder, F. (2025). An Intelligent Tutor for Planning in Large Partially Observable Environments. International Journal of Artificial Intelligence in Education, 35, 3095–3127. https://doi.org/10.1007/s40593-025-00493-7
Page metadata
Field Value
type design
id cognitive-tutor-partially-observable-planning
title Intelligent cognitive tutor for planning in partially observable environments
description An intelligent tutoring system that teaches MGPO-discovered planning strategies through interactive practice with metacognitive feedback.
status review
generated by process:settle-candidates on 2026-10-28
trust unverified — no one has confirmed this page's content
sources 1 — heindrich-2025

Frontmatter is the page's machine-readable record — see the Schema & Guide.