A Faithful Explainable Adaptive Assessment Framework Using Coupled Knowledge-Behaviour Belief Modelling for Android-Based Learning Applications.

Naga Gopi Chappidi; Kumar N S L Prakash Gandham; Hanish Venkat Sundarapalli

A Faithful Explainable Adaptive Assessment Framework Using Coupled Knowledge-Behaviour Belief Modelling for Android-Based Learning Applications.

Author(s)	Mr. Naga Gopi Chappidi, Mr. Kumar N S L Prakash Gandham, Mr. Hanish Venkat Sundarapalli
Country	India
Abstract	Adaptive assessment systems increasingly combine prediction, learner-facing explanations, and automated instructional actions, but these components are often produced by separate mechanisms. This creates a risk that the explanation shown to a learner describes one reason for difficulty while the adaptive policy responds to another. This paper introduces the Faithful Explainable Adaptive Assessment Framework (FEAAF), centered on Explanation-Action Faithfulness: the instructional action selected by the system should remain consistent with the explanation presented to the learner. FEAAF uses a Coupled Knowledge-Behaviour Belief Model (CKBM) that tracks both skill mastery and behavioural signals such as time pressure and disengagement, because either signal alone may be insufficient to distinguish different sources of underperformance. When underperformance is detected, the framework considers a small set of diagnostic categories—knowledge gap, time pressure, and disengagement—and restricts ECAP (Explanation-Consistent Adaptive Policy) to actions compatible with the selected explanation. This restriction provides a structural guarantee: the selected action cannot contradict the explanation by construction. Because the framework is intended for Android and other on-device settings, mastery and behavioural estimates are maintained with a lightweight update strategy rather than a full joint representation across all possible skill combinations. The paper presents FEAAF as a conceptual and formal framework in relation to knowledge tracing, item-response testing, intelligent tutoring systems, and explainable or attribution-based educational data mining. Within the literature reviewed here, no comparable approach was identified that explicitly enforces explanation-action consistency at the action-selection stage. No implementation, dataset, or empirical evaluation is reported; the contribution is the framework and the faithfulness property it is designed to provide.
Keywords	Explainable AI, Adaptive Assessment, Learner Modelling, Diagnostic Reasoning, Android/On-Device Learning, Educational Data Mining, Intelligent Tutoring Systems.
Field	Computer > Artificial Intelligence / Simulation / Virtual Reality
Published In	Volume 17, Issue 10, October 2026
Published On	2026-10-08

View / Download PDF File

About IJTAS Fees & Payment Current Issue Publication Archive	Submit Research Paper Track Submission Status Publication Guidelines Publication Ethics	Join as a Reviewer Editors & Reviewers Get Reviewer Membership Certi.	Website/Journal Policies Usage Policy Content Policies Privacy Policy

Contact Us	Message on WhatsApp	+91-9687-182-185	editor@ijtas.com

International Journal of Technology and Applied Science

A Widely Indexed Open Access Peer Reviewed Multidisciplinary Bi-monthly Scholarly International Journal

A Faithful Explainable Adaptive Assessment Framework Using Coupled Knowledge-Behaviour Belief Modelling for Android-Based Learning Applications.

Share this