Integração de Redes Adversárias Generativas, Transformers e Aprendizado Semi-Supervisionado para Geração de Dados Tabulares Sintéticos
Resumo
Este trabalho propõe a Semi-Supervised Tabular Transformer GAN (STTGAN), uma arquitetura que integra Redes Adversárias Generativas, aprendizado semi-supervisionado e mecanismos de atenção de Transformers para gerar dados tabulares sintéticos realistas sobre Doença Renal Crônica. A abordagem permite o treinamento com poucos rótulos, mantendo a utilidade preditiva. A STTGAN é avaliada utilizando métricas quantitativas, onde resultados demonstram que a arquitetura permite gerar dados que preservam informações clínicas, mesmo com rótulos escassos.Referências
Ghosheh, G. O., Li, J., and Zhu, T. (2024). A survey of generative adversarial networks for synthesizing structured electronic health records. In ACM Computing Surveys, volume 56, pages 1–34.
Iimori, S., Naito, S., Noda, Y., Sato, H., Nomura, N., Sohara, E., Okado, T., Sasaki, S., Uchida, S., and Rai, T. (2018). Data from: Prognosis of chronic kidney disease with normal-range proteinuria: The ckd-route study. Version 1, 2018.
Jha, V., Garcia-Garcia, G., Iseki, K., Li, Z., Naicker, S., Plattner, B., Saran, R., Wang, A. J., and Yang, C. (2013). Chronic kidney disease: global dimension and perspectives. The Lancet, 382(9888):260–272.
Kang, H. Y. J., Ko, M., and Ryu, K. S. (2025). Tabular transformer generative adversarial network for heterogeneous distribution in healthcare. Scientific Reports, 15(10254).
Liu, T., Fan, J., Li, G., Tang, N., and Du, X. (2024). Tabular data synthesis with generative adversarial networks: design space and optimizations. The VLDB Journal, 33:255–280.
Schurer, L., Assunção, J. V. C., Reinheimer, I. C., de Figueiredo, C. E. P., and Silva, L. A. L. (2025). Gans and fine-tuning through transfer learning for the generation of electronic health records on chronic kidney diseases. In 2025 IEEE 38th International Symposium on Computer-Based Medical Systems (CBMS). IEEE.
Sricharan, K., Bala, R., Shreve, M., Ding, H., Kumar, S., and Sun, J. (2017). Semi-supervised conditional gans. arXiv preprint arXiv:1708.05789.
Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Łukasz Kaiser, and Polosukhin, I. (2017). Attention is all you need. arXiv preprint arXiv:1706.03762.
Yang, X., Song, Z., King, I., and Xu, Z. (2023). A survey on deep semi-supervised learning. IEEE Transactions on Knowledge and Data Engineering, 35(9):8934–8954.
Iimori, S., Naito, S., Noda, Y., Sato, H., Nomura, N., Sohara, E., Okado, T., Sasaki, S., Uchida, S., and Rai, T. (2018). Data from: Prognosis of chronic kidney disease with normal-range proteinuria: The ckd-route study. Version 1, 2018.
Jha, V., Garcia-Garcia, G., Iseki, K., Li, Z., Naicker, S., Plattner, B., Saran, R., Wang, A. J., and Yang, C. (2013). Chronic kidney disease: global dimension and perspectives. The Lancet, 382(9888):260–272.
Kang, H. Y. J., Ko, M., and Ryu, K. S. (2025). Tabular transformer generative adversarial network for heterogeneous distribution in healthcare. Scientific Reports, 15(10254).
Liu, T., Fan, J., Li, G., Tang, N., and Du, X. (2024). Tabular data synthesis with generative adversarial networks: design space and optimizations. The VLDB Journal, 33:255–280.
Schurer, L., Assunção, J. V. C., Reinheimer, I. C., de Figueiredo, C. E. P., and Silva, L. A. L. (2025). Gans and fine-tuning through transfer learning for the generation of electronic health records on chronic kidney diseases. In 2025 IEEE 38th International Symposium on Computer-Based Medical Systems (CBMS). IEEE.
Sricharan, K., Bala, R., Shreve, M., Ding, H., Kumar, S., and Sun, J. (2017). Semi-supervised conditional gans. arXiv preprint arXiv:1708.05789.
Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Łukasz Kaiser, and Polosukhin, I. (2017). Attention is all you need. arXiv preprint arXiv:1706.03762.
Yang, X., Song, Z., King, I., and Xu, Z. (2023). A survey on deep semi-supervised learning. IEEE Transactions on Knowledge and Data Engineering, 35(9):8934–8954.
Publicado
12/11/2025
Como Citar
COSTA, Rafael A.; PREGARDIER, Rafael C.; REINHEIMER, Isabel C.; POLI-DE-FIGUEIREDO, Carlos E.; SILVA, Luis A. L..
Integração de Redes Adversárias Generativas, Transformers e Aprendizado Semi-Supervisionado para Geração de Dados Tabulares Sintéticos. In: ESCOLA REGIONAL DE APRENDIZADO DE MÁQUINA E INTELIGÊNCIA ARTIFICIAL DA REGIÃO SUL (ERAMIA-RS), 1. , 2025, Porto Alegre/RS.
Anais [...].
Porto Alegre: Sociedade Brasileira de Computação,
2025
.
p. 244-247.
DOI: https://doi.org/10.5753/eramiars.2025.16622.
