HQIAT-ML: Hybrid Quantum-Inspired Adaptive Transformer with Meta-Learning for Student Dropout Prediction Risk Explanation

DOI

10.2991/978-94-6239-713-2_56How to use a DOI?

Keywords

Predicting student dropout; Explainable Artificial Intelligence; Counterfactuals; Quantum-inspired transformers; Learning analytics; OULAD dataset; Pedagogical constraints

Abstract

This paper describes the framework HQIAT-ML, the hybrid quantum- inspired adaptive transformer with meta-learning, for predicting and explaining the risk of student dropout for the Open University Learning Analytics Datasets (OULAD). Unlike traditional explainable AI methods based on SHAP and LIME, for HQIAT-ML we constructed pedagogically constrained counterfactual intended to offer minimally intrusive, pedagogically actionable recommendations for the student and the teacher. Complex dynamics of engagement and performance are addressed through the integration of quantum- inspired feature transformation and adaptive multi-head attention with multi- scale temporal feature extraction. Training employs a hybrid focal-label smoothing loss and Mix-up augmentation to mitigate class imbalance and to favor generalization. Predictive accuracy is on the order of the best published and is represented by an AUC-ROC of 0.9615 with precision and recall balanced performance on a large-scale educational datasets. The dual approach to explaining the result of the prediction moves from the purely predictive to the actionable by providing motivational feedback to the student on an individual basis and systemically to the teacher. Evaluation has shown that our proposed approach surpasses prior benchmarks, while the discussion recognizes the limitations of the interventions, possible mechanisms, and time horizons. For future research, we intend to focus on the counterfactual interpretability and the application of causal modeling and multi-objective optimization. The HQIAT- ML framework exemplifies the first-of-its-kind combination of state-of-the-art deep learning technologies and their focus on educational explainability, allowing them to not only predict but also offer actionable recommendations that enhance equity through the optimization of student retention.

Copyright

© 2026 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Abdulkadir Shehu Bichi
AU  - Jyoti Shekhawat
PY  - 2026
DA  - 2026/06/25
TI  - HQIAT-ML: Hybrid Quantum-Inspired Adaptive Transformer with Meta-Learning for Student Dropout Prediction Risk Explanation
BT  - Proceedings of the International Conference on Advances in Computing Technology and Artificial Intelligence (COMPUTATIA 2026)
PB  - Atlantis Press
SP  - 759
EP  - 774
SN  - 2589-4919
UR  - https://doi.org/10.2991/978-94-6239-713-2_56
DO  - 10.2991/978-94-6239-713-2_56
ID  - Bichi2026
ER  -