Differential Privacy of Online Distributed optimization under Adversarial Nodes

Hou, Ming; Li, Dequan; Wu, Xiongjun; Shen, Xiuyu

doi:10.23919/chicc.2019.8865820

Cited by 8 publications

(2 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the context of centralized optimization by means of recursive algorithms, differential privacy has been applied to gradient descent [15][16][17], deep learning [18], as well as federated learning [19,20]. The decentralized setting considered in this work is studied in [6,[21][22][23][24], where independent and identically distributed perturbations are added at each agent as in (8) and differential privacy is established.…”

Section: Related Workmentioning

confidence: 99%

Graph-Homomorphic Perturbations for Private Decentralized Learning

Vlaski

Sayed

2021

ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

Decentralized algorithms for stochastic optimization and learning rely on the diffusion of information through repeated local exchanges of intermediate estimates. Such structures are particularly appealing in situations where agents may be hesitant to share raw data due to privacy concerns. Nevertheless, in the absence of additional privacy-preserving mechanisms, the exchange of local estimates, which are generated based on private data can allow for the inference of the data itself. The most common mechanism for guaranteeing privacy is the addition of perturbations to local estimates before broadcasting. These perturbations are generally chosen independently at every agent, resulting in a significant performance loss. We propose an alternative scheme, which constructs perturbations according to a particular nullspace condition, allowing them to be invisible (to first order in the step-size) to the network centroid, while preserving privacy guarantees. The analysis allows for general nonconvex loss functions, and is hence applicable to a large number of machine learning and signal processing problems, including deep learning.

show abstract

Section: Related Workmentioning

confidence: 99%

Graph-Homomorphic Perturbations for Private Decentralized Learning

Vlaski

Sayed

2021

ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

show abstract

Section: Related Workmentioning

confidence: 99%

Graph-Homomorphic Perturbations for Private Decentralized Learning

Vlaski¹,

Sayed²

2020

Preprint

View full text Add to dashboard Cite

Decentralized algorithms for stochastic optimization and learning rely on the diffusion of information as a result of repeated local exchanges of intermediate estimates. Such structures are particularly appealing in situations where agents may be hesitant to share raw data due to privacy concerns. Nevertheless, in the absence of additional privacy-preserving mechanisms, the exchange of local estimates, which are generated based on private data can allow for the inference of the data itself. The most common mechanism for guaranteeing privacy is the addition of perturbations to local estimates before broadcasting. These perturbations are generally chosen independently at every agent, resulting in a significant performance loss. We propose an alternative scheme, which constructs perturbations according to a particular nullspace condition, allowing them to be invisible (to first order in the step-size) to the network centroid, while preserving privacy guarantees. The analysis allows for general nonconvex loss functions, and is hence applicable to a large number of machine learning and signal processing problems, including deep learning.

show abstract

Privacy-Preserving Distributed Optimization and Learning

Chen,

Wang

2024

Reference Module in Materials Science and Materials Engineering

View full text Add to dashboard Cite

Differential Privacy of Online Distributed optimization under Adversarial Nodes

Cited by 8 publications

References 11 publications

Graph-Homomorphic Perturbations for Private Decentralized Learning

Graph-Homomorphic Perturbations for Private Decentralized Learning

Graph-Homomorphic Perturbations for Private Decentralized Learning

Privacy-Preserving Distributed Optimization and Learning

Contact Info

Product

Resources

About