FRAdRec: A Federated Real-Time Advertising Recommendation Framework Based on User Behavior Modeling and Cloud-Native Data Infrastructure

XiangYuan He; KuangCong Liu

doi:10.63313/JCSFT.9075

Authors

XiangYuan He Chongqing University, Chongqing, China Author
KuangCong Liu Stanford University, Stanford, CA, USA Author

DOI:

https://doi.org/10.63313/JCSFT.9075

Keywords:

Federated Recommendation, Advertising Recommendation System, User Behavior Modeling, Cloud-Native Infrastructure, Real-Time Recommendation, Transformer, Deep Interest Network, Multi-Task Learning, ROI Optimization

Abstract

Advertising recommendation systems have become a core component of modern e-commerce and digital marketing platforms, where accurately capturing dynamic user preferences and supporting large-scale real-time recommendation services are critical for improving advertising effectiveness and platform revenue. However, traditional centralized recommendation frameworks often suffer from limited scalability, delayed interest modeling, insufficient optimization of business-oriented objectives, and increasing privacy risks caused by centralized user data collection. To address these challenges, this paper proposes FRAdRec, a federated real-time advertising recommendation framework based on user behavior modeling and cloud-native data infrastructure. The proposed framework integrates a Transformer-based sequential learning module with a Deep Interest Network (DIN) attention mechanism to capture both long-term and short-term user interests from click and purchase behavior sequences. Meanwhile, a federated learning strategy is introduced to enable decentralized model training without transmitting raw user data, thereby improving privacy preservation and reducing data leakage risks. Furthermore, a multi-task optimization mechanism jointly predicts click-through rate (CTR), conversion rate (CVR), and return on investment (ROI) to enhance advertising profitability. To support industrial-scale deployment, a cloud-native streaming architecture based on Kafka, Flink, Kubernetes, and Redis is designed for low-latency recommendation and online feature updating. Experimental results on the Criteo and Alibaba Taobao datasets demonstrate that FRAdRec achieves an AUC of 0.857 and reduces recommendation latency to 41 ms, outperforming several state-of-the-art recommendation models in both recommendation accuracy and real-time serving efficiency.

References

[1] Rendle S. Factorization machines[C]//2010 IEEE International conference on data mining. IEEE, 2010: 995-1000.

[2] Cheng H T, Koc L, Harmsen J, et al. Wide & deep learning for recommender systems[C]//Proceedings of the 1st workshop on deep learning for recommender systems. 2016: 7-10.

[3] Guo H, Tang R, Ye Y, et al. DeepFM: a factorization-machine based neural network for CTR prediction[J]. arXiv preprint arXiv:1703.04247, 2017.

[4] Zhou G, Zhu X, Song C, et al. Deep interest network for click-through rate prediction[C]//Proceedings of the 24th ACM SIGKDD international conference on knowledge discovery & data mining. 2018: 1059-1068.

[5] Zhou G, Mou N, Fan Y, et al. Deep interest evolution network for click-through rate prediction[C]//Proceedings of the AAAI conference on artificial intelligence. 2019, 33(01): 5941-5948.

[6] Kang W C, McAuley J. Self-attentive sequential recommendation[C]//2018 IEEE international conference on data mining (ICDM). IEEE, 2018: 197-206.

[7] Sun F, Liu J, Wu J, et al. BERT4Rec: Sequential recommendation with bidirectional encoder representations from transformer[C]//Proceedings of the 28th ACM international conference on information and knowledge management. 2019: 1441-1450.

[8] Zhang G, Zeng H, Jiang L. Uni-FinLLM: A Unified Multimodal Large Language Model with Modular Task Heads for Micro-Level Stock Prediction and Macro-Level Systemic Risk Assessment[J]. arXiv preprint arXiv:2601.02677, 2026.

[9] McMahan B, Moore E, Ramage D, et al. Communication-efficient learning of deep networks from decentralized data[C]//Artificial intelligence and statistics. Pmlr, 2017: 1273-1282.

[10] Ammad-Ud-Din M, Ivannikova E, Khan S A, et al. Federated collaborative filtering for privacy-preserving personalized recommendation system[J]. arXiv preprint arXiv:1901.09888, 2019.

[11] Chai D, Wang L, Chen K, et al. Secure federated matrix factorization[J]. IEEE Intelligent Systems, 2020, 36(5): 11-20.

[12] Warren J, Marz N. Big Data: Principles and best practices of scalable realtime data systems[M]. Simon and Schuster, 2015.

[13] Li Z, Li X, Lin X. Design and Implementation of a Platform for Business Intelligence Knowledge Mining and Graph Construction Based on Deep Learning[C]//Proceedings of the 2nd International Symposium on Integrated Circuit Design and Integrated Systems. 2025: 137-141.

[14] Li T, Li H, Zhou Y. E-commerce Sentiment Analysis Using Fine-tuned LLaMA3 Models: A QLoRA-based Approach[J]. Journal of Technology Innovation and Engineering, 2025, 1(4).

[15] Wang T, Bai Z. TailRisk-Trans: A Transformer-Based Dynamic Tail-Risk Prediction Model with Extreme-Event–Aware Attention for Financial Markets[J]. Frontiers in Business and Finance, 2026, 3(1): 172-182.

[16] Xu S, Jiang L, Gu B. Design and Validation of a Smart Neuromorphic System Architecture for Algorithmic Trading[C]//Proceedings of the 2nd International Symposium on Integrated Circuit Design and Integrated Systems. 2025: 127-136.

[17] Zhang Y, Bai Z. GenRiskNet: A GenAI-Driven Multi-Source Heterogeneous Data Fusion Framework for Financial Risk Prediction[J]. Economics and Management Innovation, 2026, 3(1): 112-121.