CQL-GNN: Coupled question-label graph neural networks for multi-label educational question classification
Abstract Accurate tagging of educational questions with multiple knowledge labels is crucial for personalized learning and resource recommendation. However, multi-label question classification faces significant challenges: labels exhibit long-tail frequency imbalance, many labels overlap semanticall...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-08-01
|
| Series: | Journal of King Saud University: Computer and Information Sciences |
| Subjects: | |
| Online Access: | https://doi.org/10.1007/s44443-025-00208-x |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849225853965172736 |
|---|---|
| author | Liwei Gao Luojia Wang |
| author_facet | Liwei Gao Luojia Wang |
| author_sort | Liwei Gao |
| collection | DOAJ |
| description | Abstract Accurate tagging of educational questions with multiple knowledge labels is crucial for personalized learning and resource recommendation. However, multi-label question classification faces significant challenges: labels exhibit long-tail frequency imbalance, many labels overlap semantically, and questions often lack detailed solution context. In this paper, we propose a novel Coupled Question-Label Graph Neural Network (CQL-GNN) framework to address these challenges. CQL-GNN models rich relationships between questions and labels as well as among labels themselves in a unified graph structure, going beyond prior methods that either treated labels independently or only modeled label-label correlations. Each question and label is represented as a node in a heterogeneous graph, enabling dynamic message passing that propagates semantic information from questions to labels and back. This coupled graph approach allows the model to capture label co-occurrence patterns and contextualize label semantics within each question’s content. We integrate pre-trained language models to encode textual features of questions and labels, and design a two-stage propagation mechanism that iteratively refines question and label representations. Experimental results on four real-world education datasets demonstrate that our method consistently outperforms state-of-the-art baselines in terms of Precision@K and F1 scores. Notably, CQL-GNN excels at disambiguating semantically similar labels and significantly enhances the prediction of rare, long-tail labels. The proposed framework provides a robust and generalizable solution for automatic tagging of educational content, with strong potential for adaptation to other multi-label text classification tasks. |
| format | Article |
| id | doaj-art-948b66ca3ff041dda91f31f1034cd6a4 |
| institution | Kabale University |
| issn | 1319-1578 2213-1248 |
| language | English |
| publishDate | 2025-08-01 |
| publisher | Elsevier |
| record_format | Article |
| series | Journal of King Saud University: Computer and Information Sciences |
| spelling | doaj-art-948b66ca3ff041dda91f31f1034cd6a42025-08-24T11:53:26ZengElsevierJournal of King Saud University: Computer and Information Sciences1319-15782213-12482025-08-0137711710.1007/s44443-025-00208-xCQL-GNN: Coupled question-label graph neural networks for multi-label educational question classificationLiwei Gao0Luojia Wang1College of Information Engineering and Artificial Intelligence, Lanzhou University of Finance and EconomicsCollege of Economics, Lanzhou University of Finance and EconomicsAbstract Accurate tagging of educational questions with multiple knowledge labels is crucial for personalized learning and resource recommendation. However, multi-label question classification faces significant challenges: labels exhibit long-tail frequency imbalance, many labels overlap semantically, and questions often lack detailed solution context. In this paper, we propose a novel Coupled Question-Label Graph Neural Network (CQL-GNN) framework to address these challenges. CQL-GNN models rich relationships between questions and labels as well as among labels themselves in a unified graph structure, going beyond prior methods that either treated labels independently or only modeled label-label correlations. Each question and label is represented as a node in a heterogeneous graph, enabling dynamic message passing that propagates semantic information from questions to labels and back. This coupled graph approach allows the model to capture label co-occurrence patterns and contextualize label semantics within each question’s content. We integrate pre-trained language models to encode textual features of questions and labels, and design a two-stage propagation mechanism that iteratively refines question and label representations. Experimental results on four real-world education datasets demonstrate that our method consistently outperforms state-of-the-art baselines in terms of Precision@K and F1 scores. Notably, CQL-GNN excels at disambiguating semantically similar labels and significantly enhances the prediction of rare, long-tail labels. The proposed framework provides a robust and generalizable solution for automatic tagging of educational content, with strong potential for adaptation to other multi-label text classification tasks.https://doi.org/10.1007/s44443-025-00208-xMulti-label text classificationGraph neural networksNatural language processingEducational question taggingLong-tail label prediction |
| spellingShingle | Liwei Gao Luojia Wang CQL-GNN: Coupled question-label graph neural networks for multi-label educational question classification Journal of King Saud University: Computer and Information Sciences Multi-label text classification Graph neural networks Natural language processing Educational question tagging Long-tail label prediction |
| title | CQL-GNN: Coupled question-label graph neural networks for multi-label educational question classification |
| title_full | CQL-GNN: Coupled question-label graph neural networks for multi-label educational question classification |
| title_fullStr | CQL-GNN: Coupled question-label graph neural networks for multi-label educational question classification |
| title_full_unstemmed | CQL-GNN: Coupled question-label graph neural networks for multi-label educational question classification |
| title_short | CQL-GNN: Coupled question-label graph neural networks for multi-label educational question classification |
| title_sort | cql gnn coupled question label graph neural networks for multi label educational question classification |
| topic | Multi-label text classification Graph neural networks Natural language processing Educational question tagging Long-tail label prediction |
| url | https://doi.org/10.1007/s44443-025-00208-x |
| work_keys_str_mv | AT liweigao cqlgnncoupledquestionlabelgraphneuralnetworksformultilabeleducationalquestionclassification AT luojiawang cqlgnncoupledquestionlabelgraphneuralnetworksformultilabeleducationalquestionclassification |