基於聯盟區塊鏈的聯邦學習網路架構｜國立清華大學博碩士論文庫

簡易檢索 / 詳目顯示

回結果列表

研究生：	陳文遠 Chen, Wen-Yuan
論文名稱：	基於聯盟區塊鏈的聯邦學習網路架構 A Federated Learning Network Architecture based on Consortium Blockchain
指導教授：	黃之浩 Huang, Chih-Hao
口試委員:	孫敏德 Sun, Min-Te 高榮駿 Kao, Jung-Chun
學位類別：	碩士 Master
系所名稱：	電機資訊學院 - 通訊工程研究所 Communications Engineering
論文出版年：	2021
畢業學年度：	110
語文別：	英文
論文頁數：	74
中文關鍵詞：	聯盟區塊鏈、Hyperledger Fabric 、去中心化人工智慧、聯盟式學習
外文關鍵詞：	Consortium blockcain, Hyperledger Fabric, Decentralized AI, Federated learning
相關次數：	點閱：3 下載：0
分享至:	分享至facebook 分享至twitter

查詢本校圖書館目錄查詢臺灣博碩士論文知識加值系統勘誤回報

我的研究提出了名為 FLNet 的系統架構，這是一個基於 Hyperledger Fabric 聯盟區塊鏈 (Corsortium Blockchain) 並且結合了聯邦學習 (Federated Learning) 的區塊鏈網路，其目的是為了讓組織或企業可以在系統上與其客戶或其他利益相關者能在不洩漏本地數據的情況下，於區塊鏈網路上共同訓練深度學習模型，以保持高度的安全及隱私。近年來區塊鏈與人工智慧成為兩個最熱門的技術，也開始有人提出了 Decentralized AI 的想法，其概念就是結合 AI 與區塊鏈技術。因為中心化數據庫的資料可能會被竄改，且資料的出處與真實性無法被保證，倘若數據遭惡意人士操縱就會導致訓練出來的 AI 之可信度下降，甚至做出錯誤的決策。Hyperledger Fabric 聯盟區塊鏈的出現很好地解決了此議題，它不僅保有了原區塊鏈的可追溯性與不可竄改性，並且還可以使用基於 Raft Protocol 的共識演算法來增進事務處理效率。此外，Fabric 中的節點之間也互相知曉彼此身分，且還能透過通道機制來維持組織間與商業事務相關之數據的隱私性和機密性。如今 AI 已被廣泛使用，人們發現其最大的阻礙是在於如何獲取大量數據，其中以高度敏感的隱私資料最難取得，因此我演示了一個聯邦學習演算法運行於聯盟區塊鏈中的流程，讓組織與使用者可以在區塊鏈中共同訓練深度學習模型並且毌須上傳本地敏感數據至區塊鏈帳本，使得所有數據保有最高的安全性及隱私性。

關鍵字 – 聯盟區塊鏈、Hyperledger Fabric、去中心化人工智慧、聯盟式學習

This work proposed a system architecture called FLNet, a blockchain network based on the Hyperledger Fabric and Federated Learning. The purpose is to allow organizations to jointly train deep learning models with their customers or other stakeholders on a consortium blockchain network without leaking local data to keep the highest security and privacy. Currently, blockchain and AI have become popular technologies. People begin to put forward decentralized AI, which is to combine blockchain and AI. Data in a centralized database may be tampered with, and the provenance and authenticity of data cannot be guaranteed. Once the above situations happen, the credibility of AI models will reduce, and even wrong decisions will be made. Fabric solved this issue well. It maintains the traceability and immutable property of blockchain and can use Raft protocol-based consensus algorithm to improve transaction processing efficiency. Users are not willing to provide private information for training models based on the privacy rule. Hence, we demonstrate a flow that runs federated learning on Hyperledger Fabric. In this way, organizations and customers can jointly train deep learning models without uploading sensitive local data to obtain better security and privacy.

Keywords - Consortium blockchain, Hyperledger Fabric, Decentralized AI, Federated learning

摘要                                                  i
Abstract                                             ii
誌謝                    　                           iii
Acknowledgments                                      iv
Contents                                              v
List of Figures                                    viii
List of Tables                                        x
Abbreviations                                        xi
Introduction                                        1
1 Motivation ....................................   1
2 Purpose .......................................   3
3 Thesis Architecture ...........................   4
Background and Related Works                        5
1 Blockchain ....................................   5
1.1 Types of Blockchain .........................   6
1.2 Smart Contracts .............................   8
1.3 Consensus Algorithm .........................   9
2 Hyperledger Fabric ............................  10
2.1 Chaincode ...................................  11
2.2 Channel Mechanism ...........................  13
2.3 Pluggable Consensus .........................  13
3 Artificial Intelligence .......................  15
4 Decentralized AI ..............................  16
5 Federated Learning ............................  17
Problem Formulation                                21
1 Federated Learning on Public Blockchain .......  21
2 Proposed FLNet System .........................  24
Simulation Environment                             26
1 Server PC .....................................  26
2 Hyperledger Fabric ............................  27
3 Docker and Docker Compose .....................  28
System Architecture                                30
1 System Overview ...............................  30
2 Hyperledger Fabric Network Operation ..........  33
2.1 Hyperledger Fabric Components ...............  33
2.2 Hyperledger Fabric Transaction Flow .........  42
3 Combine Federated Learning Algorithm ..........  49
3.1 Make Data Contribution Proposals ............  50
3.2 Train and Commit Local Models ...............  52
3.3 Aggregate into a Global Model ...............  52
Experimental Results                               58
1 Dataset .......................................  58
2 FLNet Example Network .........................  59
3 Experimental Results ..........................  63
Conclusion                                         67
1 Discussion ....................................  67
2 Future Work ...................................  68
Bibliography                                         69
                                

[1] J. Passerat-Palmbach, T. Farnan, R. Miller, M. S. Gross, H. L. Flannery, and B. Gleim, “A blockchain-orchestrated federated learning architecture for healthcare consortia,” arXiv preprint arXiv:1910.12603, 2019.
[2] S. Nakamoto, “Bitcoin: A peer-to-peer electronic cash system,” 2008. Avali- ble: https://bitcoin.org/bitcoin.pdf.
[3] Z. Zheng, S. Xie, H. Dai, X. Chen, and H. Wang, “An overview of blockchain technology: Architecture, consensus, and future trends,” in Proceedings of the 2017 IEEE International Congress on Big Data (BigData Congress), pp. 557– 564, 2017.
[4] S. Sinha, R. Bathla, et al., “Implementation of blockchain in financial sector to improve scalability,” in 2019 4th International Conference on Information Systems and Computer Networks (ISCON), pp. 144–148, IEEE, 2019.
[5] O. R. Kabi and V. N. L. Franqueira, “Blockchain-based distributed market- place,” in Proceedings of the 2018 Business Information Systems Workshops, pp. 197–210, 2018.
[6] H. Mukne, P. Pai, S. Raut, and D. Ambawade, “Land record management us- ing hyperledger fabric and ipfs,” in Proceedings of the 2019 10th International Conference on Computing, Communication and Networking Technologies (IC-CCNT), pp. 1–8, 2019.
[7] L. Zhou, L. Wang, Y. Sun, and P. Lv, “Beekeeper: A blockchain-based iot system with secure storage and homomorphic computation,” IEEE Access, vol. 6, pp. 43472–43488, 2018.
[8] T. N. Dinh and M. T. Thai, “Ai and blockchain: A disruptive integration,”Computer, vol. 51, no. 9, pp. 48–53, 2018.
[9] K. Salah, M. H. U. Rehman, N. Nizamuddin, and A. Al-Fuqaha, “Blockchain for ai: Review and open research challenges,” IEEE Access, vol. 7, pp. 10127– 10149, 2019.
[10] V. Buterin, “Ethereum whitepaper,” 2013. Available: https://ethereum. org/en/whitepaper/.
[11] J. Chen, “Know your client (kyc),” 2021. Available: https://www. investopedia.com/terms/k/knowyourclient.asp.
[12] W. Kenton, “Anti money laundering (aml),” 2021. Available: https://www. investopedia.com/terms/a/aml.asp.
[13] The Linux Foundation, HYPERLEDGER FABRIC, 2015. Available: https://www.hyperledger.org/use/fabric.
[14] S. S. Sabry, N. M. Kaittan, and I. M. Ali, “The road to the blockchain tech- nology: Concept and types,” Periodicals of Engineering and Natural Sciences, vol. 7, no. 4, pp. 1821–1832, 2019.
[15] “Bitcoin,” 2009. Available: https://bitcoin.org/en/.
[16] “Ethereum,” 2015. Available: https://ethereum.org/en/.
[17] S. Porru, A. Pinna, M. Marchesi, and R. Tonelli, “Blockchain-oriented soft- ware engineering: Challenges and new directions,” in Proceedings of the 2017 IEEE/ACM 39th International Conference on Software Engineering Compan- ion (ICSE-C), pp. 169–171, 2017.
[18] D. Guegan, “Public blockchain versus private blockhain,” halshs.archives- ouvertes.fr/halshs-01524440, 2017.
[19] “Quorum,” 2016. Available: https://consensys.net/quorum/.
[20] X. Xie, Q. Zheng, Z. Guo, Q. Wang, and X. Li, “Design and research on b2b trading platform based on consortium blockchains,” in Proceedings of the 2018 International Conference on Cloud Computing and Security (ICCCS), pp. 436–447, 2018.
[21] W. Metcalfe, “Ethereum, smart contracts, dapps.,” in Blockchain and Crypto Currency, pp. 77–93, 2020.
[22] L. M. Bach, B. Mihaljevic, and M. Zagar, “Comparative analysis of blockchain consensus algorithms,” in Proceedings of the 2018 41st International Conven- tion on Information and Communication Technology, Electronics and Micro- electronics (MIPRO), pp. 1545–1550, 2018.
[23] N. Ariffin and A. Z. Ismail, “The design and implementation of trade finance application based on hyperledger fabric permissioned blockchain platform,” in Proceedings of the 2019 International Seminar on Research of Information Technology and Intelligent Systems (ISRITI), pp. 488–493, 2019.
[24] M. Raikwar, S. Mazumdar, S. Ruj, S. S. Gupta, A. Chattopadhyay, and K.-Y. Lam, “A blockchain framework for insurance processes,” in Proceedings of the 2018 9th IFIP International Conference on New Technologies, Mobility and Security (NTMS), pp. 1–4, 2018.
[25] A. Murugan, T. Chechare, B. Muruganantham, and S. G. Kumar, “Healthcare information exchange using blockchain technology,” International Journal of Electrical and Computer Engineering (IJECE), vol. 10, no. 1, pp. 421–426, 2020.
[26] K. Gao, Y. Liu, H. Xu, and T. Han, “Design and implementation of food supply chain traceability system based on hyperledger fabric,” International Journal of Computational Science and Engineering (IJCSE), vol. 23, no. 2, pp. 185–193, 2020.
[27] D. Kirillov, V. Korkhov, V. Petrunin, M. Makarov, I. M. Khamitov, andV. Dostov, “Implementation of an e-voting scheme using hyperledger fabric permissioned blockchain,” in Proceedings of the 2019 International Conference on Computational Science and Its Applications (ICCSA), pp. 509–521, 2019.
[28] A. Iftekhar, X. Cui, M. Hassan, and W. Afzal, “Application of blockchain and internet of things to ensure tamper-proof data availability for food safety,” Journal of Food Quality, vol. 2020, 2020.
[29] N. J. Nilsson, Principles of artificial intelligence. Morgan Kaufmann, 2014.
[30] Y. LeCun, Y. Bengio, and G. Hinton, “Deep learning,” nature, vol. 521, no. 7553, pp. 436–444, 2015.
[31] J. D. Harris and B. Waggoner, “Decentralized and collaborative ai on blockchain,” in Proceedings of the 2019 IEEE International Conference on Blockchain (Blockchain), pp. 368–375, 2019.
[32] J. Koneˇcny`, H. B. McMahan, F. X. Yu, P. Richt´arik, A. T. Suresh, and D. Ba- con, “Federated learning: Strategies for improving communication efficiency,” arXiv preprint arXiv:1610.05492, 2016.
[33] K. Bonawitz, H. Eichner, W. Grieskamp, D. Huba, A. Ingerman, V. Ivanov,C. Kiddon, J. Koneˇcny`, S. Mazzocchi, H. B. McMahan, et al., “Towards federated learning at scale: System design,” arXiv preprint arXiv:1902.01046, 2019.
[34] Q. Yang, Y. Liu, T. Chen, and Y. Tong, “Federated machine learning: Con- cept and applications,” ACM Transactions on Intelligent Systems and Tech- nology (TIST), vol. 10, no. 2, pp. 1–19, 2019.
[35] J. Dreyer, M. Fischer, and R. T¨onjes, “Performance analysis of hyperledger fabric 2.0 blockchain platform,” in Proceedings of the Workshop on Cloud Continuum Services for Smart IoT Systems, pp. 32–38, 2020.
[36] “Indian liver patient records.” Available: https://www.kaggle.com/uciml/ indian-liver-patient-records.
[37] “Heart disease dataset.” Available: https://www.kaggle.com/ johnsmith88/heart-disease-dataset.

簡易檢索 / 詳目顯示

相關論文