Differentially private block coordinate descent

Riaz, Shazia; Ali, Saqib; Wang, Guojun; Anees, Asad

doi:10.1016/j.jksuci.2022.11.017

Cited by 2 publications

(2 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This intuitively limits the amount of information that model is learning from any given example. Then noise is added, so the sample is from Gaussian distribution [96] and standard deviation of C sigma. As shown in the algorithm, high parameters C and sigma can be tuned to give epsilon delta guarantees for each step of gradient descent.…”

Section: Fundamental Law Of Information Recoverymentioning

confidence: 99%

See 1 more Smart Citation

Federated learning and differential privacy in clinical health: Extensive survey

Odera¹

2023

World J. Adv. Eng. Technol. Sci.

View full text Add to dashboard Cite

Federated Learning (FL) is concept that has been adopted in medical field to analyze data in individual devices through aggregation of machine learning model in global server. It also provides data privacy being that the sampled devices are not allowed to share data among themselves. Therefore, it minimizes computation costs and privacy risks to some extent compared to conventional methods of machine learning. However, federation learning provides a different use case in health as compared to other sectors. Preservation of patients’ sensitive information such as electronic health record (EHR) when sharing data among different medical practitioners is of greatest concern. So the question is, how should FL techniques be structured in the current clinical environment where heterogeneity is the order of the day? The EU’s General Data Protection Regulation (GDPR) and Health Insurance Portability and Accountability Act of 1996 (HIPPA) regulations recommends health providers to gain authorizations from patients before sharing their private data for medical analytical progression. This leads to some bottlenecks in clinical analysis. Although attempts have been made to address some of the challenges, privacy, performance, implementation, computation and adversaries still pose some threats. This paper provides a comprehensive review that covers literature, mathematical notations, architecture, process flow, challenges and frameworks used to implement FL with respect to healthcare. Possible solutions on how to address privacy challenges in accordance with HIPPA act and GDPR is discussed. Finally, the study gives future direction of FL in clinical health and a list of practical tools to conduct analysis on patients’ data.

show abstract

Section: Fundamental Law Of Information Recoverymentioning

confidence: 99%

“…In order to track the overall gradient descent, this article studied the papers [96], [97] that describes moments accountant [95] for tracking privacy budget. First they looked at naïve analysis of adding privacy budgets [95].…”

Section: Fundamental Law Of Information Recoverymentioning

confidence: 99%

Federated learning and differential privacy in clinical health: Extensive survey

Odera¹

2023

World J. Adv. Eng. Technol. Sci.

View full text Add to dashboard Cite

show abstract

Membership inference attack on differentially private block coordinate descent

Riaz,

Ali,

Wang

et al. 2023

PeerJ Computer Science

View full text Add to dashboard Cite

The extraordinary success of deep learning is made possible due to the availability of crowd-sourced large-scale training datasets. Mostly, these datasets contain personal and confidential information, thus, have great potential of being misused, raising privacy concerns. Consequently, privacy-preserving deep learning has become a primary research interest nowadays. One of the prominent approaches adopted to prevent the leakage of sensitive information about the training data is by implementing differential privacy during training for their differentially private training, which aims to preserve the privacy of deep learning models. Though these models are claimed to be a safeguard against privacy attacks targeting sensitive information, however, least amount of work is found in the literature to practically evaluate their capability by performing a sophisticated attack model on them. Recently, DP-BCD is proposed as an alternative to state-of-the-art DP-SGD, to preserve the privacy of deep-learning models, having low privacy cost and fast convergence speed with highly accurate prediction results. To check its practical capability, in this article, we analytically evaluate the impact of a sophisticated privacy attack called the membership inference attack against it in both black box as well as white box settings. More precisely, we inspect how much information can be inferred from a differentially private deep model’s training data. We evaluate our experiments on benchmark datasets using AUC, attacker advantage, precision, recall, and F1-score performance metrics. The experimental results exhibit that DP-BCD keeps its promise to preserve privacy against strong adversaries while providing acceptable model utility compared to state-of-the-art techniques.

show abstract

Differentially private block coordinate descent

Cited by 2 publications

References 20 publications

Federated learning and differential privacy in clinical health: Extensive survey

Federated learning and differential privacy in clinical health: Extensive survey

Membership inference attack on differentially private block coordinate descent

Contact Info

Product

Resources

About