Adaptive Career Training
Research
A research concept that combines a map of prerequisite knowledge with analysis of learner explanations to identify gaps in understanding.
The question
A correct answer does not necessarily show why a learner is correct. A plausible explanation can also conceal a missing prerequisite. I am exploring whether a structured model of professional knowledge can help an AI tutor distinguish these cases.
The proposed approach
The system would represent concepts and their dependencies as a knowledge graph. An LLM would interpret the learner’s explanations against that structure, identifying possible gaps and directing teaching towards the prerequisites that need attention.
The graph constrains what must be understood; the model provides flexible dialogue and explanation. I am deliberately considering bounded professional domains where the required body of knowledge can be defined with useful completeness.
What would make it useful
The commercial case depends on better diagnosis and more focused training, not simply the ability to generate teaching material. A system that confidently misreads a learner could waste their time or conceal a gap that matters in practice.
This remains a research concept. The key evaluation question is whether its judgements about understanding agree with expert assessment and lead to useful remediation. Improved learning outcomes have not been established.