연합 학습 환경에서의 Task-Specific Adaptive Differential Privacy 메커니즘 평가 방안 연구
Study on Evaluation Method of Task-Specific Adaptive Differential Privacy Mechanism in Federated Learning Environment
0
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
연합 학습(FL)은 여러 공동 작업자 간에 분산된 모델 학습을 위한 강력한 방법론으로 부상해 데이터 공유의 필요성을 없애준다. FL은 데이터 프라이버시를 보호하는 기능으로 호평을 받고 있지만, 다양한 유형의 프라이버시 공격으로부터 자유롭지 않다. 대표적인 개인정보 보호 기술인 차분 프라이버시(DP)는 이러한 취약점에 대응하기 위해 널리 사용된다. 이 논문에서는 기존의 작업별 적응형 DP 메커니즘을 FL 환경에 적용해 성능을 평가한다. 포괄적인 분석을 통해 다양한 DP 메커니즘이 공유 글로벌 모델의 성능에 미치는 영향을 평가하며, 특히 다양한 데이터 배포 및 분할 스키마에 주의를 기울인다. 이를 통해, FL에서 개인정보 보호와 유용성 간의 복잡한 상호 작용에 대한 이해를 심화하고, 성능 저하 없이 데이터를 보호할 수 있는 검증된 방법론을 제공한다.
Abstract
Federated Learning (FL) has emerged as a potent methodology for decentralized model training across multiple collaborators, eliminating the need for data sharing. Although FL is lauded for its capacity to preserve data privacy, it is not impervious to various types of privacy attacks. Differential Privacy (DP), recognized as the golden standard in privacy-preservation techniques, is widely employed to counteract these vulnerabilities. This paper makes a specific contribution by applying an existing, task-specific adaptive DP mechanism to the FL environment. Our comprehensive analysis evaluates the impact of this mechanism on the performance of a shared global model, with particular attention to varying data distribution and partitioning schemes. This study deepens the understanding of the complex interplay between privacy and utility in FL, providing a validated methodology for securing data without compromising performance.
참고문헌 (21)
- S. R. Pandey, "A Crowdsourcing Framework for On-Device Federated Learning", IEEE Transactions on Wireless Communications, 19(5), pp.3241~3256 (2020)
- J. Kang, "Reliable Federated Learning for Mobile Networks", IEEE Wireless Communications, 27(2), pp.72~80 (2020)
- J. Li, "A Federated Learning Based Privacy-Preserving Smart Healthcare System", IEEE Transactions on Industrial Informatics, 18(3), pp.2021~2031 (2022)
- M. Ali, "Federated Learning for Privacy Preservation in Smart Healthcare Systems : A Comprehensive Survey", IEEE Journal of Biomedical and Health Informatics, 27(2), pp.778~789 (2023)
- G. Shingi, "A federated learning based approach for loan defaults prediction", 2020 International Conference on Data Mining Workshops(ICDMW), pp.362~368 (2020)
- Ahmed Imteaj, "Leveraging asynchronous federated learning to predict customers financial distress", Intelligent Systems with Applications, 14, pp.20064~ (2022)
- Briland Hitaj, "Deep Models Under the GAN: Information Leakage from Collaborative Deep Learning", Proceedings of the 2017 ACM SIGSAC Conference on Computer and Communications Security (CCS '17), pp.603~618 (2017)
- Zhu, Ligeng, "Deep leakage from gradients", Proceedings of Neural Information Processing Systems (NIPS’19), pp.14774~14784 (2017)
- Tolpegin, Vale, "Data poisoning attacks against federated learning systems", proceedings of 25th European Symposium on Research in Computer Security, ESORICS 2020, pp.480~501 (2020)
- V. Shejwalkar, "Back to the Drawing Board: A Critical Evaluation of Poisoning Attacks on Production Federated Learning", proceedings of 2022 IEEE Symposium on Security and Privacy (SP), pp.1354~1371 (2022)
- L. Shi, "HFL-DP: Hierarchical Federated Learning with Differential Privacy", proceedings of 2021 IEEE Global Communications Conference (GLOBECOM), pp.1~7 (2021)
- K. Wei, "Federated Learning With Differential Privacy: Algorithms and Performance Analysis", IEEE Transactions on Information Forensics and Security, 15, pp.3454~3469 (2020)
- Park June Beom, "An Implementation of Federated Learning based on Blockchain", The Korean Journal of BigData, 5(1), pp.89~96 (2020)
- R. Kanagavelu, "Two-Phase Multi-Party Computation Enabled Privacy-Preserving Federated Learning", proceedings of 20th IEEE, 2020, 410~419
- Y. Li, "Privacy-Preserving Federated Learning Framework Based on Chained Secure Multiparty Computing", IEEE Internet of Things Journal, 8(8), pp.6178~6186 (2021)
- Utaliyeva A, "Task-Specific Adaptive Differential Privacy Method for Structured Data", Sensors MDPI, 23(4), pp.1980~1992 (2023)
- "Adult. UCI Machine Learning Repository" (1996)[URL]
- Scott M. Lundberg, "A Unified Approach to Interpreting Model Predictions", proceedings of Neural Information Processing Systems (NIPS’17), pp.4766~4777 (2017)
- Ribeiro, Marco Tulio, "Why should i trust you? Explaining the predictions of any classifier", Proceedings of the 22nd ACM SIGKDD, pp.1135~1144 (2016)
- [URL]