Citation:

Varun Chivukula, 2022. "The Use of Federated Learning for Digital Advertising Measurement", ESP Journal of Engineering & Technology Advancements, 2(4): 161-162.

Abstract:

Federated learning (FL) allows collaborative model training across decentralized devices or datasets without the need for raw data sharing, preserving user privacy. This paper explores FL's application in randomized control trial (RCT) measurement for digital advertising and other domains, emphasizing privacy-preserving techniques. We present a theoretical framework, demonstrate its implementation, and analyze its advantages, limitations, and recommendations.

References:

[1] McMahan, H. B., Ramage, D., et al. (2017). "Communication-Efficient Learning of Deep Networks from Decentralized Data." Proceedings of AISTATS.

[2] Dwork, C., & Roth, A. (2014). "The Algorithmic Foundations of Differential Privacy." Foundations and Trends in Theoretical Computer Science.

[3] Hard, A., Rao, K., et al. (2018). "Federated Learning for Mobile Keyboard Prediction." arXiv:1811.03604.

[4] Bonawitz, K., et al. (2019). "Practical Secure Aggregation for Privacy-Preserving Machine Learning." ACM CCS.

[5] Pearl, J. (2009). "Causality: Models, Reasoning, and Inference." Cambridge University Press.

Keywords:

Federated Learning, Randomized Control Trials, Privacy Enhancing Technologies, Digital Advertising.