Evaluating GNN Architecture Embeddings in Conjunction with Self-Supervised Training Objectives for the Session Disambiguation Process

• Matheus H. C. Leme Leme UEM https://orcid.org/0000-0002-0874-8780
• Giovani O. Santos Santos UEM https://orcid.org/0009-0001-3626-3043
• Vinicius S. Simm Simm UEM https://orcid.org/0009-0005-2541-938X
• Douglas R. Tanno Tanno UEM https://orcid.org/0000-0002-9523-1615
• Mateus Macorini Alves Alves UEM https://orcid.org/0000-0002-7995-3669
• Marcos A. Domingues UEM https://orcid.org/0000-0001-7195-0714

DOI: https://doi.org/10.5753/eniac.2026.30085

Resumo

Noisy interaction histories in recommender systems can degrade personalization, posing a structural challenge analogous to the shared accounts problem. In this work, we extract and evaluate latent representations using four GNN architectures under four self-supervised learning techniques, combined with hybridization with a sequential attention model. An empirical evaluation of 16 configurations across two domains reveals that standalone graph-based approaches suffer from representational collapse in subjective scenarios. However, the hybridization can act as an essential topological corrector, suppressing noise-induced similarities and precisely isolating the user’s current intent, thereby contributing to persona disambiguation.

Referências

Akiba, T., Sano, S., Yanase, T., Ohta, T., and Koyama, M. (2019). Optuna: A nextgeneration hyperparameter optimization framework. In Proceedings of the 25th ACM SIGKDD international conference on knowledge discovery & data mining, pages 2623–2631.

Belkin, M. and Niyogi, P. (2003). Laplacian eigenmaps for dimensionality reduction and data representation. Neural computation, 15(6):1373–1396.

Hamilton, W., Ying, Z., and Leskovec, J. (2017). Inductive representation learning on large graphs. Advances in neural information processing systems, 30.

He, X., Deng, K., Wang, X., Li, Y., Zhang, Y., and Wang, M. (2020). Lightgcn: Simplifying and powering graph convolution network for recommendation. In Proceedings of the 43rd International ACM SIGIR conference on research and development in Information Retrieval, pages 639–648.

Kang, W.-C. and McAuley, J. (2018). Self-attentive sequential recommendation. In 2018 IEEE International Conference on Data Mining (ICDM), pages 197–206. IEEE.

Kipf, T. N. and Welling, M. (2016a). Semi-supervised classification with graph convolutional networks.

Kipf, T. N. and Welling, M. (2016b). Variational graph auto-encoders.

Ma, M., Ren, P., Lin, Y., Chen, Z., Ma, J., and Rijke, M. d. (2019). π-net: A parallel information-sharing network for shared-account cross-domain sequential recommendations. In Proceedings of the 42nd international ACM SIGIR conference on research and development in information retrieval, pages 685–694.

Peintner, A., Moscati, M., Parada-Cabaleiro, E., Schedl, M., and Zangerle, E. (2025). Unsupervised graph embeddings for session-based recommendation with item features.

Santana, I. A. P., Pinhelli, F., Donini, J., Catharin, L., Mangolin, R. B., Feltrim, V. D., Domingues, M. A., et al. (2020). Music4all: A new music database and its applications. In 2020 International Conference on Systems, Signals and Image Processing (IWSSIP), pages 399–404.

Veličković, P., Fedus, W., Hamilton, W. L., Liò, P., Bengio, Y., and Hjelm, R. D. (2018). Deep graph infomax.

Wan, Z., Liu, X., Wang, B., Qiu, J., Li, B., Guo, T., Chen, G., and Wang, Y. (2023). Spatio-temporal contrastive learning-enhanced gnns for session-based recommendation. ACM Transactions on Information Systems, 42(2):1–26.

Wang, F., Gao, X., Chen, Z., and Lyu, L. (2023). Contrastive multi-level graph neural networks for session-based recommendation. IEEE Transactions on Multimedia, 25:9278–9289.

Wang, S., Cao, L., Wang, Y., Sheng, Q. Z., Orgun, M. A., and Lian, D. (2021). A survey on session-based recommender systems. ACM Computing Surveys (CSUR), 54(7):1–38.

Wen, X., Peng, Z., Huang, S., Wang, S., and Yu, P. S. (2021). Miss: A multi-user identification network for shared-account session-aware recommendation. In International Conference on Database Systems for Advanced Applications, pages 228–243.

Wu, S., Tang, Y., Zhu, Y., Wang, L., Xie, X., and Tan, T. (2019). Session-based recommendation with graph neural networks. In Proceedings of the AAAI conference on artificial intelligence, volume 33, pages 346–353.

Xu, K., Hu, W., Leskovec, J., and Jegelka, S. (2018). How powerful are graph neural networks? Ying, R., He, R., Chen, K., Eksombatchai, P., Hamilton, W. L., and Leskovec, J. (2018). Graph convolutional neural networks for web-scale recommender systems. In Proceedings of the 24th ACM SIGKDD international conference on knowledge discovery & data mining, pages 974–983.

Zhang, A., Fawaz, N., Ioannidis, S., and Montanari, A. (2012). Guess who rated this movie: Identifying users through subspace clustering. arXiv preprint arXiv:1208.1544.

Zhang, J., Zhao, Z., Li, C., and Yu, Y. (2025). Lightweight yet fine-grained: A graph capsule convolutional network with subspace alignment for shared-account sequential recommendation. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 39, pages 13242–13250.

Zhu, Y., Xu, Y., Yu, F., Liu, Q., Wu, S., and Wang, L. (2020). Deep graph contrastive representation learning.

Zykov, R., Noskov, A., and Anokhin, A. (2022). Retailrocket recommender system dataset.