The ubiquity of financial APIs has generated new levels of security risk, and traditional centralized fraud detection systems exhibit significant limitations in distributed financial environments, such as data privacy constraints and compliance with real-time response requirements. To address these gaps, this work proposes a novel federated learning architecture dedicated to efficient real-time financial API attack detection by combining state-of-the-art differential privacy techniques with cross-institutional cooperation protocols. The approach provides a distributed topology design for secure collaboration of multiple financial service providers satisfying strict data sovereignty requirements and implementing streaming API call data processing pipelines and low-latency federation model updates via asynchronous aggregation protocols. Experimental results show significant performance gains compared to traditional centralized systems, namely improving detection accuracy from 86.8% to 93.2% detection accuracy with the federated learning approach, as well as keeping the processing latencies well below 10 milliseconds and diminishing the false positive rate to 2.3%. The privacy protection measures ensure data leakage probability remains below 0.003%, meeting stringent financial regulatory requirements of financial control due to the mathematically provable privacy bounds. Comprehensive evaluation across six well-defined attack categories that demonstrates better detection capabilities, especially on the detection of sophisticated attack forms that need multimodal institution specific views. Financial mid-scale institution networks can benefit both economically and from better security performance, primarily: (a) 78% improvement in cost- effectiveness, with a simply economic approach; and (b) it also presents an annual saving of $2.4 million. The study adds new hybridization of differential privacy with real time threat detection to the current literature, which provides a new level of maturity for theoretical (computational) analysis and practical implementation of cooperative cyber security frameworks, and also demonstrates efficient arguments for when and how federated learning should be used for enterprise service based on financial APIs.

Luqing Ren. Real-time Threat Identification Systems for Financial API Attacks under Federated Learning Framework. Academic Journal of Business & Management (2025), Vol. 7, Issue 10: 65-71. https://doi.org/10.25236/AJBM.2025.071010.

[1] Jianping, W., Guangqiu, Q., Chunming, W., Weiwei, J., & Jiahe, J. (2024). Federated learning for network attack detection using attention-based graph neural networks. Scientific Reports, 14(1), 19088.

[2] Karunamurthy, A., Vijayan, K., Kshirsagar, P. R., & Tan, K. T. (2025). An optimal federated learning-based intrusion detection for IoT environment. Scientific Reports, 15(1), 8696.

[3] Khraisat, A., Alazab, A., Singh, S., Jan, T., & Jr. Gomez, A. (2024). Survey on federated learning for intrusion detection system: Concept, architectures, aggregation strategies, challenges, and future directions. ACM Computing Surveys, 57(1), 1-38.

[4] Liu, J., Lou, J., Xiong, L., Liu, J., & Meng, X. (2024, December). Cross-silo federated learning with record-level personalized differential privacy. In Proceedings of the 2024 on ACM SIGSAC Conference on Computer and Communications Security (pp. 303-317).

[5] He, P., Lin, C., & Montoya, I. (2024). DPFedBank: Crafting a Privacy-Preserving Federated Learning Framework for Financial Institutions with Policy Pillars. arXiv preprint arXiv:2410.13753.

[6] Basu, P., Roy, T. S., Naidu, R., & Muftuoglu, Z. (2021). Privacy enabled financial text classification using differential privacy and federated learning. arXiv preprint arXiv:2110.01643.

[7] Hernandez-Ramos, J., Karopoulos, G., Chatzoglou, E., Kouliaridis, V., Marmol, E., Gonzalez-Vidal, A., & Kambourakis, G. (2023). Intrusion Detection based on Federated Learning: a systematic review. ACM Computing Surveys.

[8] Mahmud, S. A., Islam, N., Islam, Z., Rahman, Z., & Mehedi, S. T. (2024). Privacy-preserving federated learning-based intrusion detection technique for cyber-physical systems. Mathematics, 12(20), 3194.

[9] Xia, Z., & Saha, S. C. (2025). FinGraphFL: Financial Graph-Based Federated Learning for Enhanced Credit Card Fraud Detection. Mathematics, 13(9), 1396.

[10] Shenoy, D., Bhat, R., & Krishna Prakasha, K. (2025). Exploring privacy mechanisms and metrics in federated learning. Artificial Intelligence Review, 58(8), 223.