IEEE 2986-2023

$44.42

IEEE Recommended Practice for Privacy and Security for Federated Machine Learning (Approved Draft)

Published By	Publication Date	Number of Pages
IEEE	2023	57

Guaranteed Safe Checkout

Category: IEEE

If you have any questions, feel free to reach out to our online customer service team by clicking on the bottom right corner. We’re here to assist you 24/7.
Email:[email protected]

Description

New IEEE Standard – Active. Privacy and security issues pose great challenges to the federated machine leaning (FML) community. A general view on privacy and security risks while meeting applicable privacy and security requirements in FML is provided. This recommended practice is provided in four parts: malicious failure and non-malicious failure in FML, privacy and security requirements from the perspective of system and FML participants, defensive methods and fault recovery methods, and the privacy and security risks evaluation. It also provides some guidance for typical FML scenarios in different industry areas, which can facilitate practitioners to use FML in a better way.

PDF Catalog

PDF Pages	PDF Title
1	IEEE Std 2986-2023 Front Cover
2	Title page
4	Important Notices and Disclaimers Concerning IEEE Standards Documents
8	Participants
10	Introduction
11	Contents
12	1. Overview 1.1 Scope 1.2 Purpose
13	1.3 Word usage 2. Normative references 3. Definitions, acronyms, and abbreviations 3.1 Definitions
14	3.2 Acronyms and abbreviations
15	4. Common failures 4.1 Non-malicious failure 4.1.1 Participant reporting failure 4.1.2 Noisy model updates 4.1.3 Non-IID
16	4.2 Malicious failure 4.2.1 Data attacks 4.2.1.1 Training sample sniffing 4.2.1.2 Recovery of training data properties
17	4.2.2 Model attacks 4.2.2.1 Before training
18	4.2.2.2 During training
19	5. Privacy and security requirements 5.1 From the perspective of system 5.1.1 Confidentiality 5.1.1.1 Identity authentication 5.1.1.2 Authorization
20	5.1.1.3 Interface security 5.1.1.4 Transmission 5.1.1.5 Cryptographic algorithm 5.1.2 Integrity 5.1.3 Availability
21	5.1.4 Controllability 5.1.5 Robustness 5.1.6 Privacy-preserving
22	5.2 From the perspective of FML participants 5.2.1 Data owner
23	5.2.2 Coordinator 5.2.3 Model user
24	5.2.4 Auditor 6. Defensive methods and fault recovery methods 6.1 Fault recovery methods for non-malicious failure
25	6.2 Defensive methods for data attack 6.2.1 Secure multiparty computation
26	6.2.2 Differential privacy 6.2.3 AI-based approaches
27	6.2.4 Other methods
29	6.3 Defensive methods for model attacks 6.3.1 Adversarial training 6.3.2 Malicious participants detection
31	6.3.3 Resilient aggregation 6.3.4 Digital signature for model
33	7. Evaluation 7.1 Privacy 7.1.1 Privacy risk identification
35	7.1.2 Privacy risk analysis
36	7.1.3 Privacy risk evaluation:
37	7.2 Security 7.2.1 Security risk identification
38	7.2.2 Security risk analysis
39	7.2.3 Security risk evaluation
41	Annex A (normative) Example of a native privacy-preserving method
42	Annex B (normative) Example of a privacy protection method of member inference attacks
43	Annex C (informative) Example of the quantitative evaluation method
44	Annex D (informative) FML scenarios D.1 FML in distributed manufacturing management scenario D.1.1 Data owner
45	D.1.2 Coordinator D.1.3 Model user
46	D.2 FML in financial scenario D.2.1 Basic requirements D.2.2 Financial data security requirements
47	D.2.3 Web network security requirements D.2.4 Communication security requirements D.2.5 Verification and audit
48	D.2.6 Password requirements
49	Annex E (informative) Example of calculating the existing probability
50	Annex F (informative) Risk level for privacy and security F.1 Risk level for privacy
51	F.2 Risk level for security
54	Annex G (informative) Bibliography
57	Back Cover

Additional information

Standard Title	IEEE Recommended Practice for Privacy and Security for Federated Machine Learning (Approved Draft)
Published Code	IEEE
Publication Date	2023
Pages Count	57

Preview PDF


PDF Format	Searchable	Printable	Preview Now

IEEE 2986-2023

PDF Catalog

Related products