پیوندهای مفید
آمار
| تعداد نشریات | 32 |
| تعداد شماره‌ها | 604 |
| تعداد مقالات | 5,885 |
| تعداد مشاهده مقاله | 8,899,430 |
| تعداد دریافت فایل اصل مقاله | 6,524,677 |
An Efficient Hybrid GCN-DDPG Model for Node Classification in Large-Scale Graphs | ||
| Contributions of Science and Technology for Engineering | ||
| دوره 3، شماره 3، آبان 2026، صفحه 61-74 اصل مقاله (974.97 K) | ||
| نوع مقاله: Original Article | ||
| شناسه دیجیتال (DOI): 10.22080/cste.2026.31137.1115 | ||
| نویسنده | ||
| Fatemeh Fathi Nezhad* | ||
| Department of Computer Engineering and Information Technology, Razi University, Kermanshah Province, Iran | ||
| تاریخ دریافت: 08 بهمن 1404، تاریخ بازنگری: 30 اردیبهشت 1405، تاریخ پذیرش: 14 آبان 1405 | ||
| چکیده | ||
| Graph-structured data present significant challenges for machine learning due to their non-Euclidean nature, complex topological dependencies, and heterogeneous node features. This paper proposes a hybrid framework that integrates graph convolutional networks (GCNs) with the deep deterministic policy gradient (DDPG) algorithm to improve node classification in large-scale graphs. Unlike conventional GCNs with fixed feature aggregation, the proposed model employs a reinforcement learning agent based on continuous-action representations, which operates on GCN-generated node embeddings and learns a policy that maps these embeddings to class preference vectors using a reward signal. To enhance efficiency and stability, a random forest-based feature selection mechanism is incorporated, and a neighborhood-aware reward function is designed to enforce label consistency among connected nodes. Experiments conducted on five benchmark datasets Cora, Citeseer, Pubmed, Twitch-Gamers, and FacebookPagePage demonstrate that the proposed GCN-DDPG model consistently outperforms baseline methods, including GCN, GAT, and other GCN–reinforcement learning variants. The proposed framework achieves relative accuracy improvements of +11.91% on Pubmed, +10.84% on Citeseer, +9.06% on Cora, and +5.57% on Twitch compared to the baseline GCN model, while exhibiting low variance across repeated runs, indicating strong robustness. Overall, the results confirm that continuous-control reinforcement learning, when properly aligned with graph neural architectures and reward design, provides an effective and scalable solution for high-fidelity node classification in complex graph-structured data. | ||
| کلیدواژه‌ها | ||
| Graph Convolutional Networks؛ Deep Reinforcement Learning؛ Deep Deterministic Policy Gradient؛ Node Classification؛ Neighborhood Consistency | ||
| مراجع | ||
|
[1] K. C. Chuang, P. S. Cheng, Y. H. Tsai, and M. H. Tsai, “Establishing a GRU-GCN coordination-based prediction model for miRNA-disease associations,” BMC Genomic Data, vol. 26, no. 1, p. 4, 2025, doi:10.1186/s12863-024-01293-z.
[2] X. Wang, Z. Zhang, G. Shen, S. Lai, Y. Chen, and S. Zhu, “Multi-view knowledge graph convolutional networks for recommendation,” Appl. Soft Comput., vol. 169, p. 112633, 2025, doi:10.1016/j.asoc.2024.112633.
[3] J. Liu, S. Chen, and L. Shen, “A comprehensive survey on graph neural network accelerators,” Front. Comput. Sci., vol. 19, no. 2, p. 192104, 2025, doi:10.1007/s11704-023-3307-2.
[4] L. Deng, X. Zhang, L. J. Slater, H. Liu, and S. Tao, “Integrating Euclidean and non-Euclidean spatial information for deep learning-based spatiotemporal hydrological simulation,” J. Hydrol., vol. 638, p. 131438, 2024, doi:10.1016/j.jhydrol.2024.131438.
[5] Q. Dang, Q. Liu, S. Yang, and X. He, “Data-Driven Evolutionary Algorithm Based on Inductive Graph Neural Networks for Multimodal Multiobjective Optimization,” IEEE Trans. Evol. Comput., vol. 30, no. 1, pp. 186–198, 2026, doi:10.1109/TEVC.2025.3541046.
[6] J. Gilmer, S. S. Schoenholz, P. F. Riley, O. Vinyals, and G. E. Dahl, “Message Passing Neural Networks,” in Lecture Notes in Physics, vol. 968, 2020, pp. 199–214. doi:10.1007/978-3-030-40245-7_10.
[7] T. N. Kipf and M. Welling, “Semi-supervised classification with graph convolutional networks,” in 5th International Conference on Learning Representations, ICLR 2017 - Conference Track Proceedings, 2017.
[8] S. Ghode and M. Digalwar, “Enhanced Electric Vehicle charging strategy through Graph Convolutional Networks integrated with deep reinforcement learning,” Int. J. Inf. Technol., vol. 17, no. 7, pp. 4217–4229, 2025, doi:10.1007/s41870-024-02394-9.
[9] T. Cui, X. Yang, F. Jia, J. Jin, Y. Ye, and R. Bai, “Mobile robot sequential decision making using a deep reinforcement learning hyper-heuristic approach,” Expert Systems with Applications, vol. 257, p. 124959, Dec. 2024, doi: 10.1016/j.eswa.2024.124959.
[10] V. Mnih et al., “Asynchronous methods for deep reinforcement learning,” 33rd International Conference on Machine Learning, ICML, New York, United States, 19-24 June, 2016, pp. 2850–2869.
[11] S. Hochreiter and J. Schmidhuber, “Long short-term memory,” Neural Comput., vol. 9, no. 8, pp. 1735–1780.
[12] T. Wang, “CLE: An Integrated Framework of CNN, LSTM, and Enhanced A3C for Addressing Multi-Agent Pathfinding Challenges in Warehousing Systems,” IEEE Access, vol. 12, pp. 88904–88912, 2024, doi:10.1109/access.2024.3416111.
[13] L. Tian, W. Wu, and T. Yu, “Graph Random Forest: A Graph Embedded Algorithm for Identifying Highly Connected Important Features,” Biomolecules, vol. 13, no. 7, p. 1153, Jul. 2023, doi: 10.3390/biom13071153.
[14] M. Sewak, “Deep Q network (DQN), double DQN, and dueling DQN,” in Deep Reinforcement Learning. Springer, Singapore, 2019, doi:10.1007/978-981-13-8285-7_8.
[15] Z. Wang, T. Schaul, M. Hessel, H. van Hasselt, M. Lanctot, and N. de Freitas, “Dueling network architectures for deep reinforcement learning,” in Proc. Int. Conf. Mach. Learn. (ICML), New York, United States, 19-24 June, 2016, pp. 1995–2003.
[16] J. Leng et al., “Deep Reinforcement Learning of Graph Convolutional Neural Network for Resilient Production Control of Mass Individualized Prototyping Toward Industry 5.0,” IEEE Transactions on Systems, Man, and Cybernetics: Systems, vol. 54, no. 11, pp. 7092–7105, Nov. 2024, doi:10.1109/tsmc.2024.3446671.
[17] P. Rokhforoz, M. Montazeri, and O. Fink, “Multi-agent reinforcement learning with graph convolutional neural networks for optimal bidding strategies of generation units in electricity markets,” Expert Syst. Appl., vol. 225, p. 120010, 2023, doi:10.1016/j.eswa.2023.120010.
[18] X. Gao, X. Li, Q. Liu, Z. Li, F. Yang, and T. Luan, “Multi-Agent Decision-Making Modes in Uncertain Interactive Traffic Scenarios via Graph Convolution-Based Deep Reinforcement Learning,” Sensors, vol. 22, no. 12, p. 4586, 2022, doi: 10.3390/s22124586.
[19] H. Weng, Y. Hu, M. Liang, J. Xi, and B. Yin, “Optimizing bidding strategy in electricity market based on graph convolutional neural network and deep reinforcement learning,” Appl. Energy, vol. 380, p. 124978, 2025, doi:10.1016/j.apenergy.2024.124978.
[20] H. Wu, L. Tian, H. Tang, R. Li, and P. Jiao, “Graph Convolutional Reinforcement Learning-Guided Joint Trajectory Optimization and Task Offloading for Aerial Edge Computing,” IEEE Transactions on Intelligent Transportation Systems, vol. 26, no. 10, pp. 17487–17498, Oct. 2025, doi:10.1109/tits.2024.3490533.
[21] G. Nurbek, “Exploring graph neural networks in reinforcement learning: A comparative study on architectures for locomotion tasks,” Master’s thesis, The University of Texas Rio Grande Valley, Edinburg, United States, 2024.
[22] S. Ding, W. Du, L. Ding, J. Zhang, L. Guo, and B. An, “Multiagent Reinforcement Learning with Graphical Mutual Information Maximization,” IEEE Trans. Neural Networks Learn. Syst., vol. 36, no. 11, pp. 19515–19524, 2025, doi:10.1109/TNNLS.2023.3243557.
[23] Y. Guo, Q. Wu, and H. She, “A Routing Optimization Policy Using Graph Convolution Deep Reinforcement Learning,” in 2023 IEEE/CIC International Conference on Communications in China, ICCC 2023, 2023, pp. 1–6. doi:10.1109/ICCC57788.2023.10233329.
[24] J. Chen and Z. Wu, “Dynamic Computation Offloading with Energy Harvesting Devices: A Graph-Based Deep Reinforcement Learning Approach,” IEEE Commun. Lett., vol. 25, no. 9, pp. 2968–2972, 2021, doi:10.1109/LCOMM.2021.3094842.
[25] T. Peng et al., “GraphRARE: Reinforcement Learning Enhanced Graph Neural Network with Relative Entropy,” in Proceedings - International Conference on Data Engineering, 2024, pp. 2489–2502. doi:10.1109/ICDE60146.2024.00196.
[26] M. Valko, R. Munos, B. Kveton, and T. Kocák, “Spectral bandits for smooth graph functions,” in Proc. 31st Int. Conf. Mach. Learn. (ICML), Beijing, China, 21–26, Jun, 2014, pp. 46–54.
[27] M. Zhang, Z. Cui, M. Neumann, and Y. Chen, “An End-to-End Deep Learning Architecture for Graph Classification,” Proceedings of the AAAI Conference on Artificial Intelligence, vol. 32, no. 1, Apr. 2018, doi:10.1609/aaai.v32i1.11782. | ||
|
آمار تعداد مشاهده مقاله: 3 |
||