Task2Vec Readiness: Diagnostics for Federated Learning from Pre-Training Embeddings
Resumo
O desempenho da Aprendizagem Federada (FL) é altamente sensível à heterogeneidade dos clientes, mas os usuários carecem de ferramentas para antecipar os resultados do treinamento antes que ele comece. Introduzimos o Task2Vec Readiness, um framework de diagnóstico pré-treinamento que embute o conjunto de dados local de cada cliente em um vetor numérico e deriva índices escalares de prontidão a partir do espaço de embeddings da federação. Em quatro conjuntos de dados, três tamanhos de federação e dois algoritmos de agregação, as correlações de Pearson e Spearman entre os índices propostos e o desempenho final frequentemente superam 0,9 e permanecem significativas após correção para múltiplas comparações, estabelecendo a geometria da federação como um proxy do sucesso do treinamento colaborativo que não exige acesso aos dados brutos dos clientes nem conhecimento do processo que gerou a partição.Referências
Achille, A., Lam, M., Tewari, R., Ravichandran, A., Maji, S., Fowlkes, C. C., Soatto, S., and Perona, P. (2019). Task2vec: Task embedding for meta-learning. In Proceedings of the IEEE/CVF international conference on computer vision, pages 6430–6439.
Albaseer, A., Abdallah, M., Al-Fuqaha, A., and Erbad, A. (2021). Client selection approach in support of clustered federated learning over wireless edge networks.
Caldas, S., Duddu, S. M. K., Wu, P., Li, T., Konečnỳ, J., McMahan, H. B., Smith, V., and Talwalkar, A. (2018). Leaf: A benchmark for federated settings. arXiv preprint arXiv:1812.01097.
He, K., Zhang, X., Ren, S., and Sun, J. (2016). Deep residual learning for image recognition. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 770–778.
Krizhevsky, A., Hinton, G., et al. (2009). Learning multiple layers of features from tiny images.
Li, T., Sahu, A. K., Zaheer, M., Sanjabi, M., Talwalkar, A., and Smith, V. (2020). Federated optimization in heterogeneous networks. Proceedings of Machine learning and systems, 2:429–450.
Lutz, A., Steidl, G., Müller, K., and Samek, W. (2025). Optimizing federated learning by entropy-based client selection. In 2025 3rd International Conference on Federated Learning Technologies and Applications (FLTA), pages 218–227.
McMahan, B., Moore, E., Ramage, D., Hampson, S., and Arcas, B. A. y. (2017). Communication-Efficient Learning of Deep Networks from Decentralized Data. In Singh, A. and Zhu, J., editors, Proceedings of the 20th International Conference on Artificial Intelligence and Statistics, volume 54 of Proceedings of Machine Learning Research, pages 1273–1282. PMLR.
Orlandi, F. C., Dos Anjos, J. C. S., Leithardt, V. R. Q., De Paz Santana, J. F., and Geyer, C. F. R. (2023). Entropy to mitigate non-iid data problem on federated learning for the edge intelligence environment. IEEE Access, 11:78845–78857.
Pei, J., Liu, W., Li, J., Wang, L., and Liu, C. (2024). A review of federated learning methods in heterogeneous scenarios. IEEE Transactions on Consumer Electronics, 70(3):5983–5999.
PyTorch (2025). torchvision.models.resnet34. [link]. Accessed: 2025-09-23.
Solans, D., Heikkila, M., Vitaletti, A., Kourtellis, N., Anagnostopoulos, A., Chatzigiannakis, I., et al. (2024). Non-iid data in federated learning: A survey with taxonomy, metrics, methods, frameworks and future directions. arXiv preprint arXiv:2411.12377.
Xu, S., Xia, H., Zhang, R., Liu, P., and Fu, Y. (2024). Fednor: A robust training framework for federated learning based on normal aggregation. Information Sciences, 684:121274.
Yang, J., Shi, R., Wei, D., Liu, Z., Zhao, L., Ke, B., Pfister, H., and Ni, B. (2023). Medmnist v2-a large-scale lightweight benchmark for 2d and 3d biomedical image classification. Scientific Data, 10(1):41.
Yang, Z., Zhang, Y., Li, C., Cheung, Y.-m., Han, B., and Yuan, Y. (2025). Fedgps: Statistical rectification against data heterogeneity in federated learning. In Belgrave, D., Zhang, C., Lin, H., Pascanu, R., Koniusz, P., Ghassemi, M., and Chen, N., editors, Advances in Neural Information Processing Systems, volume 38, pages 94731–94772. Curran Associates, Inc.
Zhang, C., Xie, Y., Bai, H., Yu, B., Li, W., and Gao, Y. (2021). A survey on federated learning. Knowledge-Based Systems, 216:106775.
Albaseer, A., Abdallah, M., Al-Fuqaha, A., and Erbad, A. (2021). Client selection approach in support of clustered federated learning over wireless edge networks.
Caldas, S., Duddu, S. M. K., Wu, P., Li, T., Konečnỳ, J., McMahan, H. B., Smith, V., and Talwalkar, A. (2018). Leaf: A benchmark for federated settings. arXiv preprint arXiv:1812.01097.
He, K., Zhang, X., Ren, S., and Sun, J. (2016). Deep residual learning for image recognition. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 770–778.
Krizhevsky, A., Hinton, G., et al. (2009). Learning multiple layers of features from tiny images.
Li, T., Sahu, A. K., Zaheer, M., Sanjabi, M., Talwalkar, A., and Smith, V. (2020). Federated optimization in heterogeneous networks. Proceedings of Machine learning and systems, 2:429–450.
Lutz, A., Steidl, G., Müller, K., and Samek, W. (2025). Optimizing federated learning by entropy-based client selection. In 2025 3rd International Conference on Federated Learning Technologies and Applications (FLTA), pages 218–227.
McMahan, B., Moore, E., Ramage, D., Hampson, S., and Arcas, B. A. y. (2017). Communication-Efficient Learning of Deep Networks from Decentralized Data. In Singh, A. and Zhu, J., editors, Proceedings of the 20th International Conference on Artificial Intelligence and Statistics, volume 54 of Proceedings of Machine Learning Research, pages 1273–1282. PMLR.
Orlandi, F. C., Dos Anjos, J. C. S., Leithardt, V. R. Q., De Paz Santana, J. F., and Geyer, C. F. R. (2023). Entropy to mitigate non-iid data problem on federated learning for the edge intelligence environment. IEEE Access, 11:78845–78857.
Pei, J., Liu, W., Li, J., Wang, L., and Liu, C. (2024). A review of federated learning methods in heterogeneous scenarios. IEEE Transactions on Consumer Electronics, 70(3):5983–5999.
PyTorch (2025). torchvision.models.resnet34. [link]. Accessed: 2025-09-23.
Solans, D., Heikkila, M., Vitaletti, A., Kourtellis, N., Anagnostopoulos, A., Chatzigiannakis, I., et al. (2024). Non-iid data in federated learning: A survey with taxonomy, metrics, methods, frameworks and future directions. arXiv preprint arXiv:2411.12377.
Xu, S., Xia, H., Zhang, R., Liu, P., and Fu, Y. (2024). Fednor: A robust training framework for federated learning based on normal aggregation. Information Sciences, 684:121274.
Yang, J., Shi, R., Wei, D., Liu, Z., Zhao, L., Ke, B., Pfister, H., and Ni, B. (2023). Medmnist v2-a large-scale lightweight benchmark for 2d and 3d biomedical image classification. Scientific Data, 10(1):41.
Yang, Z., Zhang, Y., Li, C., Cheung, Y.-m., Han, B., and Yuan, Y. (2025). Fedgps: Statistical rectification against data heterogeneity in federated learning. In Belgrave, D., Zhang, C., Lin, H., Pascanu, R., Koniusz, P., Ghassemi, M., and Chen, N., editors, Advances in Neural Information Processing Systems, volume 38, pages 94731–94772. Curran Associates, Inc.
Zhang, C., Xie, Y., Bai, H., Yu, B., Li, W., and Gao, Y. (2021). A survey on federated learning. Knowledge-Based Systems, 216:106775.
Publicado
19/10/2026
Como Citar
MAFUZ, Cristiano; SILVA, Pedro; SILVA, Rodrigo.
Task2Vec Readiness: Diagnostics for Federated Learning from Pre-Training Embeddings. In: ENCONTRO NACIONAL DE INTELIGÊNCIA ARTIFICIAL E COMPUTACIONAL (ENIAC), 23. , 2026, Cuiabá/MT.
Anais [...].
Porto Alegre: Sociedade Brasileira de Computação,
2026
.
p. 2943-2954.
ISSN 2763-9061.
DOI: https://doi.org/10.5753/eniac.2026.31005.
