Cognitive Agent Compilation for Explicit Problem Solver Modeling

2026-05-07

AI Summary

This work addresses the limitation of current large language models in educational settings, where knowledge representations are neither verifiable nor editable, hindering explicit modeling of learners’ knowledge states, misconceptions, and problem-solving strategies. To overcome this, the paper introduces the Cognitive Agent Compilation (CAC) framework, which, for the first time, integrates cognitive architecture principles into educational AI. Leveraging a strong teacher large language model, CAC compiles problem-solving knowledge into structured goal-directed agents that explicitly separate knowledge representation, solution strategies, and verification-update mechanisms. This approach enables human-in-the-loop intervention and controllable reasoning. A prototype implementation demonstrates a fundamental trade-off between explicit control and generalization capability, laying the groundwork for interpretable, bounded-knowledge educational AI systems.

Abstract

Large language models (LLMs) are widely used for tutoring, feedback generation, and content creation, but their broad pretraining makes them hard to constrain and poor substitutes for controllable learners. Educational systems often require inspectable and editable knowledge states: educators want to know what a system assumes the learner knows, and learners benefit when the system can justify actions in terms of explicit skills, misconceptions, and strategies. Inspired by cognitive architectures, we propose Cognitive Agent Compilation (CAC), a framework that uses a strong teacher LLM to compile problem-solving knowledge into an explicit target agent. CAC separates (i) knowledge representation, (ii) problem-solving policy, and (iii) verification and update rules, with the goal of making bounded problem solving more inspectable and editable in educational settings. We present an early proof of concept implemented with Small Language Models that surfaces key design trade-offs, particularly between explicit control and scalable generalization, and positions CAC as an initial step toward bounded-knowledge AI for educational applications.

Problem

• cognitive agent
• explicit knowledge
• educational AI
• inspectable systems
• bounded problem solving

Innovation

• Cognitive Agent Compilation
• explicit knowledge representation
• inspectable AI
• educational AI
• bounded-knowledge systems