Federated Feature Learning for Secure Cross-Organizational Data Science Workflows
DOI:
https://doi.org/10.15662/IJEETR.2024.0601006Keywords:
Data Science, Federated Learning, Privacy Preserving Machine Learning, The Cross-Organizational CollaborationsAbstract
In cross-organizational data science, the premises tend to increase prediction performance, which is given by data collection. Although there are restrictions on data sharing due to privacy, security, and regulatory considerations, federal learning may remove the limits that allow collaborative model training without centralizing data communication.
The paper introduces a federal feature that uses learning as a secure technique to accept cross-organizational data in the science workflow. Therefore, establishing a federated future learning framework requires the organization to collaboratively learn common feature representations while maintaining people's data at the location level. In this context, the simulation results and enterprise-inspired scenarios typically demonstrate demonstrations of an approach that improves the utility model, maintains data security, and supports and secures inter-organizational cooperation.
Furthermore, this study focuses on federated features that demonstrate learning rather than determining conventional federated model aggregation as a feature representation and as a kind of organizational foundations that enable downstream analytics across enterprises. In reality, by offering learning representations without revealing raw data, we will be proposing a strategy that is securing collaboration while also retaining the organization's autonomy and regulatory compliance. This paradigm is particularly well-suited to providing the enterprise environment with data heterogeneity, privacy constraints, and even competition limits that restrict centralized data exchange. The study will focus on federated features learning through the simulation process and enterprise inspiration scenarios, which provide a healthy and practical balance between predictive utility and data protection, resulting in viable solutions that are scalable and secure across organizational data science workflows.
References
1. Roy, A. G., Siddiqui, S., Pölsterl, S., Navab, N., & Wachinger, C. (2019). Braintorrent: A peer-to-peer environment for decentralized federated learning. arXiv preprint arXiv:1905.06731.
https://arxiv.org/abs/1905.06731
2. Johnston, N., Eban, E., Gordon, A., & Ballé, J. (2019). Computationally efficient neural image compression. arXiv preprint arXiv:1912.08771.
https://arxiv.org/abs/1912.08771
3. Zheleva, M., Bogdanov, P., Zois, D. S., Xiong, W., Chandra, R., & Kimball, M. (2017, June). Smallholder agriculture in the information age: Limits and opportunities. In Proceedings of the 2017 Workshop on Computing within Limits (pp. 59-70).
https://dl.acm.org/doi/abs/10.1145/3080556.3080563
4. Daudt, R. C., Le Saux, B., Boulch, A., & Gousseau, Y. (2019). Multitask learning for large-scale semantic change detection. Computer Vision and Image Understanding, 187, 102783.
5. Wang, K., Mathews, R., Kiddon, C., Eichner, H., Beaufays, F., & Ramage, D. (2019). Federated evaluation of on-device personalization. arXiv preprint arXiv:1910.10252.
https://arxiv.org/abs/1910.10252
6. Yao, X., Huang, T., Zhang, R. X., Li, R., & Sun, L. (2019). Federated learning with unbiased gradient aggregation and controllable meta updating. arXiv preprint arXiv:1910.08234.
https://arxiv.org/abs/1910.08234
7. Luping, W. A. N. G., Wei, W. A. N. G., & Bo, L. I. (2019, July). CMFL: Mitigating communication overhead for federated learning. In 2019, IEEE 39th International Conference on Distributed Computing Systems (ICDCS) (pp. 954-964). IEEE.
