Tune3: Otimização Multiestágio de Hiperparâmetros com Redução de Custo e Maior Estabilidade em Modelos de Detecção de Malware Android
Resumo
O Tune3 é um método adaptativo multiestágio para otimizar hiperparâmetros em modelos neurais para cibersegurança. Ele integra inicialização informada, amostragem de extremos e refinamento iterativo, reduzindo o custo e o número de execuções necessárias. Em testes com conjuntos Android, o Tune3 iguala ou supera o Random Search com até 44% menos tempo de busca, mantendo estabilidade e qualidade na geração e classificação de dados.Referências
Basri, A., Hassan, A., and Yusof, N. (2023). A hyperparameter tuning framework for tabular synthetic data generation methods. JCVIS, 5(1).
Bergstra, J. and Bengio, Y. (2012). Random search for hyper-parameter optimization. The journal of machine learning research, 13(1).
Bischl, B., Binder, M., Lang, M., Pielok, T., Richter, J., Coors, S., Thomas, J., Ullmann, T., Becker, M., Boulesteix, A.-L., et al. (2023). Hyperparameter optimization: Foundations, algorithms, best practices, and open challenges. WIRE, 13(2).
Esteban, C., Hyland, S. L., and Rätsch, G. (2017). Real-valued (medical) time series generation with recurrent conditional gans. arXiv preprint arXiv:1706.02633.
Goodfellow, I. J., Pouget-Abadie, J., Mirza, M., Xu, B., Warde-Farley, D., Ozair, S., Courville, A., and Bengio, Y. (2014). Generative adversarial nets. Advances in neural information processing systems, 27.
Huang, W. and et al. (2020). Malware detection using deep learning techniques: A survey. ACM Computing Surveys, 53(4):1–36.
Li, L., Jamieson, K., DeSalvo, G., Rostamizadeh, A., and Talwalkar, A. (2021). A survey of hyperparameter optimization in deep learning. IEEE Access, 9:125676–125694.
Li, Q., Zhang, Y., and Wang, J. (2022). Tts-cgan: Time-series conditional generative adversarial network for tabular data synthesis. arXiv preprint arXiv:2206.13676.
Liao, L., Li, H., Shang, W., and Ma, L. (2022). An empirical study of the impact of hyperparameter tuning and model optimization on the performance properties of deep neural networks. ACM TOSEM, 31(3).
Mirza, M. and Osindero, S. (2014). Conditional generative adversarial nets. arXiv pre-print arXiv:1411.1784.
Nogueira, A., Paim, K., Bragança, H., Mansilha, R., and Kreutz, D. (2024). MalSyn-Gen: redes neurais artificiais na geração de dados tabulares sintéticos para detecção de malware. In Anais do XXIV SBSeg. SBC.
Nurhayati, R., Sari, A., and Pratama, M. (2021). Bitcoin prediction model using deep learning with hyperparameter optimization. In 2021 9th ICIM.
Ozkan-Okay, M., Akin, E., Aslan, Ö., Kosunalp, S., Iliev, T., Stoyanov, I., and Beloev, I. (2024). A comprehensive survey: Evaluating the efficiency of artificial intelligence and machine learning techniques on cyber security solutions. IEEe Access, 12.
Xu, L., Skoularidou, M., Cuesta-Infante, A., and Veeramachaneni, K. (2024). Rigorous experimental analysis of tabular data generated using ctgan and tvae. IJACSA, 15(4).
Zhou, K., Zhang, Y., and Wu, X. (2020). Time series forecasting and classification models using deep learning and attention mechanisms. Sensors, 20(24):7211.
Bergstra, J. and Bengio, Y. (2012). Random search for hyper-parameter optimization. The journal of machine learning research, 13(1).
Bischl, B., Binder, M., Lang, M., Pielok, T., Richter, J., Coors, S., Thomas, J., Ullmann, T., Becker, M., Boulesteix, A.-L., et al. (2023). Hyperparameter optimization: Foundations, algorithms, best practices, and open challenges. WIRE, 13(2).
Esteban, C., Hyland, S. L., and Rätsch, G. (2017). Real-valued (medical) time series generation with recurrent conditional gans. arXiv preprint arXiv:1706.02633.
Goodfellow, I. J., Pouget-Abadie, J., Mirza, M., Xu, B., Warde-Farley, D., Ozair, S., Courville, A., and Bengio, Y. (2014). Generative adversarial nets. Advances in neural information processing systems, 27.
Huang, W. and et al. (2020). Malware detection using deep learning techniques: A survey. ACM Computing Surveys, 53(4):1–36.
Li, L., Jamieson, K., DeSalvo, G., Rostamizadeh, A., and Talwalkar, A. (2021). A survey of hyperparameter optimization in deep learning. IEEE Access, 9:125676–125694.
Li, Q., Zhang, Y., and Wang, J. (2022). Tts-cgan: Time-series conditional generative adversarial network for tabular data synthesis. arXiv preprint arXiv:2206.13676.
Liao, L., Li, H., Shang, W., and Ma, L. (2022). An empirical study of the impact of hyperparameter tuning and model optimization on the performance properties of deep neural networks. ACM TOSEM, 31(3).
Mirza, M. and Osindero, S. (2014). Conditional generative adversarial nets. arXiv pre-print arXiv:1411.1784.
Nogueira, A., Paim, K., Bragança, H., Mansilha, R., and Kreutz, D. (2024). MalSyn-Gen: redes neurais artificiais na geração de dados tabulares sintéticos para detecção de malware. In Anais do XXIV SBSeg. SBC.
Nurhayati, R., Sari, A., and Pratama, M. (2021). Bitcoin prediction model using deep learning with hyperparameter optimization. In 2021 9th ICIM.
Ozkan-Okay, M., Akin, E., Aslan, Ö., Kosunalp, S., Iliev, T., Stoyanov, I., and Beloev, I. (2024). A comprehensive survey: Evaluating the efficiency of artificial intelligence and machine learning techniques on cyber security solutions. IEEe Access, 12.
Xu, L., Skoularidou, M., Cuesta-Infante, A., and Veeramachaneni, K. (2024). Rigorous experimental analysis of tabular data generated using ctgan and tvae. IJACSA, 15(4).
Zhou, K., Zhang, Y., and Wu, X. (2020). Time series forecasting and classification models using deep learning and attention mechanisms. Sensors, 20(24):7211.
Publicado
08/12/2025
Como Citar
OLIVEIRA, Lucas Ferreira Areias de; DINIZ, Angelo Gaspar; KREUTZ, Diego; SCHMIDT, Dionatan; MANSILHA, Rodrigo.
Tune3: Otimização Multiestágio de Hiperparâmetros com Redução de Custo e Maior Estabilidade em Modelos de Detecção de Malware Android. In: ESCOLA REGIONAL DE REDES DE COMPUTADORES (ERRC), 22. , 2025, Porto Alegre/RS.
Anais [...].
Porto Alegre: Sociedade Brasileira de Computação,
2025
.
p. 165-171.
DOI: https://doi.org/10.5753/errc.2025.17793.
