Mao Chen
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Mao Chen has a total of 29 co-authors. Here are the top co-authors by number of shared publications:
Xicheng Peng
12 shared publicationsJingzhong Zhang
5 shared publicationsXiangyang Tang
Faculty of Artificial Intelligence in Education Central China Normal University Wuhan 430079 China
4 shared publicationsChao Wu
3 shared publicationsSannyuya Liu
3 shared publicationsYajing Yang
Faculty of Artificial Intelligence in Education, Central China Normal University, Wuhan, China
3 shared publications
Latest Publications
- Open Access
- 13 November 2026
- Published by Springer Nature in Complex & Intelligent Systems
A two-stage adaptive beam search framework for personalized school timetabling under complex curriculum reform constraints
- 13 November 2026
- Published by Springer Nature in Complex & Intelligent Systems
Open AccessA two-stage adaptive beam search framework for personalized school timetabling under complex curriculum reform constraints
- 3 August 2026
- Published by Springer Nature in Journal of Automated Reasoning
- Vol. 70 (2)
Geometric Interpretation and Application of Non-degenerate Conditions in Geometric Theorem Proving
- 3 August 2026
- Published by Springer Nature in Journal of Automated Reasoning
- Vol. 70 (2)
Geometric Interpretation and Application of Non-degenerate Conditions in Geometric Theorem Proving
- 4 March 2026
- Published by SAGE Publications in Journal of Information Science
The inductive knowledge graph completion method based on neighbourhood relational learning and BERT
Inductive reasoning in knowledge graphs excels at predicting relationships between previously unseen entities, a task that transductive methods often fail to address effectively. Most current inductive approaches to relationship prediction in knowledge graphs depend on closed subgraphs for inference. However, these methods face significant limitations, as an overdependence on closed subgraphs can result in the omission of valuable neighbourhood relational information, especially in sparse subgraphs. This limitation hinders the effective utilisation of available information. To address this issue, we propose a novel method that incorporates adjacent relationships – those directly linked to the target triple – and extracts first-order adjacent relationship subgraphs centred on the target triple. The extracted information, which includes the target triple and its direct contextual relationships from both the head and tail entities, is then input into a pre-trained BERT model for further fine-tuning. Our empirical findings indicate that the approach surpasses other state-of-the-art models when evaluated on the FB15K-237 and WN18RR datasets. Specifically, compared with the SiaILP hybrid model introduced in 2024, our BERTNR model shows an average improvement of 3.9% in the AUC-PR metric and 9.63% in the Hits@10 metric across both datasets. These results underscore the effectiveness of integrating neighbourhood relational learning with BERT for inductive relationship prediction, especially for sparse graphs.
- 4 March 2026
- Published by SAGE Publications in Journal of Information Science
The inductive knowledge graph completion method based on neighbourhood relational learning and BERT
Inductive reasoning in knowledge graphs excels at predicting relationships between previously unseen entities, a task that transductive methods often fail to address effectively. Most current inductive approaches to relationship prediction in knowledge graphs depend on closed subgraphs for inference. However, these methods face significant limitations, as an overdependence on closed subgraphs can result in the omission of valuable neighbourhood relational information, especially in sparse subgraphs. This limitation hinders the effective utilisation of available information. To address this issue, we propose a novel method that incorporates adjacent relationships – those directly linked to the target triple – and extracts first-order adjacent relationship subgraphs centred on the target triple. The extracted information, which includes the target triple and its direct contextual relationships from both the head and tail entities, is then input into a pre-trained BERT model for further fine-tuning. Our empirical findings indicate that the approach surpasses other state-of-the-art models when evaluated on the FB15K-237 and WN18RR datasets. Specifically, compared with the SiaILP hybrid model introduced in 2024, our BERTNR model shows an average improvement of 3.9% in the AUC-PR metric and 9.63% in the Hits@10 metric across both datasets. These results underscore the effectiveness of integrating neighbourhood relational learning with BERT for inductive relationship prediction, especially for sparse graphs.
- 15 December 2025
- Published by Elsevier in Applied Soft Computing
- Vol. 188
Memory-based iterated local search with multiple perturbation operators for personalized learning path planning
- 15 December 2025
- Published by Elsevier in Applied Soft Computing
- Vol. 188
Memory-based iterated local search with multiple perturbation operators for personalized learning path planning
- 21 October 2025
- Published by Elsevier in Applied Soft Computing
- Vol. 186
A two-stage symbolic regression method for discovering mathematical formulas
- 21 October 2025
- Published by Elsevier in Applied Soft Computing
- Vol. 186
A two-stage symbolic regression method for discovering mathematical formulas
- 29 July 2025
- Published by Springer Nature in Annals of Mathematics and Artificial Intelligence
- Vol. 94 (5) , 751-772
Heuristic strategies for geometry theorem extension based on complex number identity*
- 29 July 2025
- Published by Springer Nature in Annals of Mathematics and Artificial Intelligence
- Vol. 94 (5) , 751-772
Heuristic strategies for geometry theorem extension based on complex number identity*
- 2 March 2025
- Published by Wiley in International Transactions in Operational Research
- Vol. 32 (6) , 3729-3755
Filtered beam search algorithm for the two‐dimensional rectangular packing problem
In this study, a filtered beam search (FBS) algorithm is proposed to address the two‐dimensional rectangle packing problem, whose objective is to minimize the height of a rectangular sheet of fixed width in which n rectangular items can be packed without overlapping. The corner‐occupying placement (COP) strategy is introduced in the packing process, which packs each rectangle into the sheet so that it occupies a corner position formed by the sheet's edges and the edges of already packed rectangles. Two evaluation criteria, namely, local evaluation and global evaluation, are proposed to evaluate the benefit of a candidate COP and are seamlessly integrated into the FBS framework in order to maintain a better balance between the solution quality and computation time. Extensive testing on four sets of 112 zero‐waste test instances indicated that the proposed method competes favourably with the state‐of‐the‐art reference methods. Furthermore, the proposed method is modified to handle a variant of the rectangle packing problem where item rotations are not allowed. The modified algorithm is competitive, compared with the existing top algorithms, according to computational results on widely used benchmarks.
- 2 March 2025
- Published by Wiley in International Transactions in Operational Research
- Vol. 32 (6) , 3729-3755
Filtered beam search algorithm for the two‐dimensional rectangular packing problem
In this study, a filtered beam search (FBS) algorithm is proposed to address the two‐dimensional rectangle packing problem, whose objective is to minimize the height of a rectangular sheet of fixed width in which n rectangular items can be packed without overlapping. The corner‐occupying placement (COP) strategy is introduced in the packing process, which packs each rectangle into the sheet so that it occupies a corner position formed by the sheet's edges and the edges of already packed rectangles. Two evaluation criteria, namely, local evaluation and global evaluation, are proposed to evaluate the benefit of a candidate COP and are seamlessly integrated into the FBS framework in order to maintain a better balance between the solution quality and computation time. Extensive testing on four sets of 112 zero‐waste test instances indicated that the proposed method competes favourably with the state‐of‐the‐art reference methods. Furthermore, the proposed method is modified to handle a variant of the rectangle packing problem where item rotations are not allowed. The modified algorithm is competitive, compared with the existing top algorithms, according to computational results on widely used benchmarks.
- 8 July 2024
- Published by Springer Nature in The Journal of Supercomputing
- Vol. 80 (16) , 24781-24800
DEKGCI: A double-ended recommendation model for integrating knowledge graph and user–item interaction graph
- 8 July 2024
- Published by Springer Nature in The Journal of Supercomputing
- Vol. 80 (16) , 24781-24800
DEKGCI: A double-ended recommendation model for integrating knowledge graph and user–item interaction graph
- 20 November 2023
- Published by Springer Nature in Journal of Automated Reasoning
- Vol. 67 (4) , 1-18
Self-evident Automated Geometric Theorem Proving Based on Complex Number Identity
- 20 November 2023
- Published by Springer Nature in Journal of Automated Reasoning
- Vol. 67 (4) , 1-18
Self-evident Automated Geometric Theorem Proving Based on Complex Number Identity
- 21 July 2023
- Published by Elsevier in Information Sciences
- Vol. 646
Task-driven cleaning and pruning of noisy knowledge graph
- 21 July 2023
- Published by Elsevier in Information Sciences
- Vol. 646
Task-driven cleaning and pruning of noisy knowledge graph