Learner digital twin links risk alerts to learning evidence
An Incremental Knowledge-Graph Learner Digital Twin for Evidence-Grounded Early Warning
Artificial Intelligence
Summary
Sometimes, tools that warn teachers a student might fail don’t explain why they think so. This paper introduces a digital twin, like a smart model of a learner, that keeps updating with new learning records and shows exactly which events or actions triggered an alert. The authors use a knowledge graph that connects data points, predictions, and explanations in one place. Their approach keeps the accuracy of usual models but adds clear, traceable reasons for each warning.
What this means in practice
- •For learning platform developers: Integrate evidence-grounded early-warning alerts that link risk predictions to specific learning events and assessment data for better intervention transparency.
- •For educational data engineers: Build analytics workflows that transform raw learning and assessment traces into interpretable graph-based features supporting consistent risk prediction over time.
Authors
Ngoc Luyen Le, Marie-Hélène Abel, Bertrand Laforge
Abstract
Early-warning models in learning analytics can identify learners at risk of failure or withdrawal, but the resulting alerts are often only weakly connected to inspectable learning evidence. A risk score may indicate that intervention is needed without showing which traces, assessment events, or temporal patterns justify the alert. This paper introduces an incremental Learner Digital Twin for evidence-grounded early warning, implemented as a lightweight knowledge graph. The twin is updated at temporal checkpoints by replaying learning traces and assessment records, and it serves both as a feature-extraction substrate and as an explanation-grounding substrate. Predictive features are derived through graph-mediated queries, while features, attributions, alerts, evidence items, and raw records are represented as queryable graph entities. The framework is instantiated and evaluated on a public learning-analytics benchmark under a temporal protocol. The results show that graph-mediated feature extraction preserves the predictive behavior of equivalent raw-table features across model families and checkpoints. Generated alerts are materialized as evidence chains linking risk decisions to feature attributions, evidence items, and raw platform records. The study shows that early-warning prediction can be routed through a semantic and provenance-aware representation while preserving predictive performance and adding traceability, verifiability, and evidence-grounded explanation.