[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-123703-en":3,"doc-seo-123703-105":30,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":4,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":27,"seo_description":14,"update_tm":28,"read_time":29},123703,34359740700684,"Finn","https://ap-avatar.wpscdn.com/avatar/1f400023980c374ae676?_k=1777273430885731487",8,"Research & Report","Privacy of 5G Enabled Networks - Homomorphic Encryption based Privacy-Preserving Machine Learning","Homomorphic encryption enables computations directly on encrypted data, supporting privacy and anonymity for sensitive information and enabling multiple parties to share computation without exposing raw inputs. Federated learning further improves privacy by avoiding centralized data sharing, yet model-parameter exchange can still leak information through inference attacks. This thesis studies federated learning with homomorphic encryption to enhance privacy in 5G networks, showing minimal degradation in predictive performance while substantially increasing computation time, with reported overhead depending on the number of clients.","| \u003Cbr>FACULTY OF SCIENCE AND TECHNOLOGY\u003Cbr>MASTER’S THESIS |  |\n| --- | --- |\n| Study programme / specialisation: Data Science | The spring semester, 2023\u003Cbr>Open / ~~Confidential~~ |\n| Author: Emil Alan Pierzgalski |  |\n| Supervisor at UiS: Ferhat Özgur Catak |  |\n| Thesis title:\u003Cbr>Privacy of 5G Enabled Networks: Homomorphic Encryption based PrivacyPreserving Machine Learning |  |\n| Credits (ECTS): 30 |  |\n| Keywords:\u003Cbr>Spectrum Sensing, Machine Learning, Federated Learning, Homomorphic Encryption | Pages: 31\u003Cbr>+ appendix: Link to repository\u003Cbr>Stavanger, July 15, 2023 |\n\nAbstract  \nHomomorphic encryption (HE) is a technique that allows computations to be performed on encrypted data, just as if the data were unencrypted. This has numerous potential applications, such as sensitive medical data, mainly when privacy and anonymity are critical. HE can also be used in cases where multiple parties need to perform computations on shared data without revealing the data to one another. One fascinating application of HE is in machine learning, specifically in a process known as federated learning (FL) . FL is a cutting-edge method that is particularly useful in situations where privacy is essential, as it eliminates the need for data to be shared with a central server, as is the case with traditional distributed machine learning models. However, privacy risks are associated with sharing model parameters, as inference attacks can obtain sensitive information. This issue can be addressed by encrypting the model parameters with HE on the client side and aggregating the encrypted data. In this paper, we explore federated learning with homomorphic encryption to improve the privacy of 5G networks. The results of our experiments show that encryption has a minimal effect on the the predictive performance of the model, but leads to an increase in computation time by 587 %, 624 % and 679 % for 2, 5, and 7 clients, respectively.  \nAcknowledgements  \nI would like to express my gratitude to my supervisor, Ferhat Özgur Catak, for introducing me to the main topics of the project, as well as his continuous guidance and excellent feedback along the way.  \nContents  \nAbstract i  \nAcknowledgements ii  \n1 Introduction 1  \n1.1 Background ........................... 1  \n1.2 Problem Statement ....................... 2  \n1.3 Outline and Contribution .................... 2  \n2 Theory 3  \n2.1 Homomorphic Encryption .................... 3  \n2.1.1 CKKS Scheme ...................... 4  \n2.2 Convolutional Neural Networks ................. 6  \n2.3 Federated Learning ....................... 7  \n2.4 Spectrum Sensing ........................ 8  \n3 Methodology 9  \n3.1 Data ............................... 10  \n3.2 Tools and materials ....................... 11  \n3.3 Solution Approach ........................ 11  \n3.3.1 Model Selection ..................... 11  \n3.3.2 Pyfhel .......................... 13  \n3.3.3 Flower .......................... 14  \n3.3.4 Strategy ......................... 15  \n3.3.5 Weighting function ................... 16  \n3.3.6 Serialization and Deserialization ............. 16  \n3.3.7 Encryption ........................ 17  \n3.3.8 Aggregation ....................... 18  \n3.3.9 Decryption ........................ 18  \n3.4 Evaluation metrics ........................ 19  \n3.4.1 Loss ........................... 19  \n3.4.2 Accuracy ......................... 20  \n3.4.3 Intersection over Union ................. 20  \n3.4.4 F-Score ......................... 21  \n3.5 Limitations and Challenges ................... 21  \n4 Results 22  \n4.1 Baseline model .......................... 22  \n4.2 FL Without Encryption ..................... 23  \n4.3 FL With Encryption ....................... 24  \n4.4 Security levels .......................... 26  \n4.5 Scale parameter ......................... 27  \n5 Discussion 29  \n6 Conclusion 30  \nBibliography 35  \nA Additional Information 36  \nA.1 GitHub repository ........................ 36  \nAcronyms  \nCKKS Cheong-Kim-Kim-Song","cbCais5RgFuq8Wke","https://ap.wps.com/l/cbCais5RgFuq8Wke","pdf",2262174,1,43,"English","en",105,"# Abstract\n# Acknowledgements\n# 1 Introduction\n## 1.1 Background\n## 1.2 Problem Statement\n## 1.3 Outline and Contribution\n# 2 Theory\n## 2.1 Homomorphic Encryption\n## 2.2 Convolutional Neural Networks\n## 2.3 Federated Learning\n## 2.4 Spectrum Sensing\n# 3 Methodology\n## 3.1 Data\n## 3.3 Solution Approach\n## 3.4 Evaluation metrics\n## 3.5 Limitations and Challenges\n# 4 Results\n## 4.1 Baseline model\n## 4.2 FL Without Encryption\n## 4.3 FL With Encryption\n## 4.4 Security levels\n## 4.5 Scale parameter\n# 5 Discussion\n# 6 Conclusion\n# Bibliography\n# A Additional Information\n## A.1 GitHub repository","[{\"question\":\"What problem does the thesis address in 5G-related machine learning?\",\"answer\":\"It addresses privacy risks when applying federated learning in 5G networks, where sharing model parameters can allow inference attacks to reveal sensitive information.\"},{\"question\":\"How does homomorphic encryption help protect federated learning model parameters?\",\"answer\":\"It encrypts model parameters on the client side so aggregation can be performed on encrypted values, preventing direct exposure of parameters to other parties.\"},{\"question\":\"What impact does encryption have on predictive performance and computation time?\",\"answer\":\"Experiments show minimal effect on predictive performance, while computation time increases substantially, with overhead reported for different client counts.\"}]","Privacy of 5G Enabled Networks - Homomorphic Encryption based Privacy-Preserving Machine Learning | PDF",1785818096,108,{"code":4,"msg":31,"data":32},"ok",{"site_id":24,"language":23,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"privacy-of-5g-enabled-networks-homomorphic-encryption-based-privacy-preserving-machine-learning","",{"@graph":36,"@context":85},[37,54,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":43,"position":53},"https://docshare.wps.com/document/privacy-of-5g-enabled-networks-homomorphic-encryption-based-privacy-preserving-machine-learning/123703/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":23,"description":14,"dateModified":62,"datePublished":62,"encodingFormat":61,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-04",true,{"@type":65,"interactionType":66,"userInteractionCount":4},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What problem does the thesis address in 5G-related machine learning?","Question",{"text":75,"@type":76},"It addresses privacy risks when applying federated learning in 5G networks, where sharing model parameters can allow inference attacks to reveal sensitive information.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does homomorphic encryption help protect federated learning model parameters?",{"text":80,"@type":76},"It encrypts model parameters on the client side so aggregation can be performed on encrypted values, preventing direct exposure of parameters to other parties.",{"name":82,"@type":73,"acceptedAnswer":83},"What impact does encryption have on predictive performance and computation time?",{"text":84,"@type":76},"Experiments show minimal effect on predictive performance, while computation time increases substantially, with overhead reported for different client counts.","https://schema.org",{"og:url":52,"og:type":87,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":89,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":46,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":46,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":46,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]