Differential Privacy and Machine Learning: a Survey and Review

Ji, Zhanglong; Elkan, Charles

doi:10.48550/arxiv.1412.7584

Cited by 63 publications

(57 citation statements)

References 50 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Privacy is a significant concern in all systems where the use of personal data has significant social and economic impact [29]. Concerns over privacy in AI systems are particularly prevalent, with the high volume of data used for sensitive decisions, such as advertisement, surveillance, health-care decisions, and money lending [14].…”

Section: Privacymentioning

confidence: 99%

Know Your Model (KYM): Increasing Trust in AI and Machine Learning

Roszel,

Norvill,

Hilger

et al. 2021

Preprint

View full text Add to dashboard Cite

The widespread utilization of AI systems has drawn attention to the potential impacts of such systems on society. Of particular concern are the consequences that prediction errors may have on realworld scenarios, and the trust humanity places in AI systems. It is necessary to understand how we can evaluate trustworthiness in AI and how individuals and entities alike can develop trustworthy AI systems. In this paper, we analyze each element of trustworthiness and provide a set of 20 guidelines that can be leveraged to ensure optimal AI functionality while taking into account the greater ethical, technical, and practical impacts to humanity. Moreover, the guidelines help ensure that trustworthiness is provable and can be demonstrated, they are implementation agnostic, and they can be applied to any AI system in any sector.

show abstract

Section: Privacymentioning

confidence: 99%

Know Your Model (KYM): Increasing Trust in AI and Machine Learning

Roszel,

Norvill,

Hilger

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…One possible choice to achieve a high quality analysis avoiding the difficulties of the centralized analysis is usage of privacy-preserving computation. There are two typical types of the privacy-preserving computation techniques based on cryptography [3,8,12,15] and differential privacy [1,4,13].…”

Section: Related Workmentioning

confidence: 99%

“…Examples to overcome the difficulties of the centralized analysis include usage of privacypreserving computation based on cryptography [3,8,12,15] and differential privacy [1,4,13]. The federated learning [16,18], that centralizes a model while the original datasets remains distributed has also been studied for this context.…”

Section: Introductionmentioning

confidence: 99%

Data collaboration analysis for distributed datasets

Imakura,

Sakurai

2019

Preprint

View full text Add to dashboard Cite

In this paper, we propose a data collaboration analysis method for distributed datasets. The proposed method is a centralized machine learning while training datasets and models remain distributed over some institutions. Recently, data became large and distributed with decreasing costs of data collection. If we can centralize these distributed datasets and analyse them as one dataset, we expect to obtain novel insight and achieve a higher prediction performance compared with individual analyses on each distributed dataset. However, it is generally difficult to centralize the original datasets due to their huge data size or regarding a privacy-preserving problem. To avoid these difficulties, we propose a data collaboration analysis method for distributed datasets without sharing the original datasets. The proposed method centralizes only intermediate representation constructed individually instead of the original dataset.

show abstract

“…Differential privacy (DP) [20] provides a rigorous privacy guarantee for data release and has become the state-of-the-art method for designing privacy-preserving ML algorithms [28]. Recently, DP has been applied to the iterative training of DNNs using stochastic gradient descent (DP-SGD) [1] and the FL algorithm of federated averaging (DP-FedAvg) [47], which have achieved competitive performances (utility) with a strong privacy guarantee.…”

Section: Introductionmentioning

confidence: 99%

Differentially Private Federated Bayesian Optimization with Distributed Exploration

Dai¹,

Low²,

Jaillet³

2021

Preprint

View full text Add to dashboard Cite

Bayesian optimization (BO) has recently been extended to the federated learning (FL) setting by the federated Thompson sampling (FTS) algorithm, which has promising applications such as federated hyperparameter tuning. However, FTS is not equipped with a rigorous privacy guarantee which is an important consideration in FL. Recent works have incorporated differential privacy (DP) into the training of deep neural networks through a general framework for adding DP to iterative algorithms. Following this general DP framework, our work here integrates DP into FTS to preserve user-level privacy. We also leverage the ability of this general DP framework to handle different parameter vectors, as well as the technique of local modeling for BO, to further improve the utility of our algorithm through distributed exploration (DE). The resulting differentially private FTS with DE (DP-FTS-DE) algorithm is endowed with theoretical guarantees for both the privacy and utility and is amenable to interesting theoretical insights about the privacy-utility trade-off. We also use real-world experiments to show that DP-FTS-DE achieves high utility (competitive performance) with a strong privacy guarantee (small privacy loss) and induces a trade-off between privacy and utility.35th Conference on Neural Information Processing Systems (NeurIPS 2021).

show abstract

Differential Privacy and Machine Learning: a Survey and Review

Cited by 63 publications

References 50 publications

Know Your Model (KYM): Increasing Trust in AI and Machine Learning

Know Your Model (KYM): Increasing Trust in AI and Machine Learning

Data collaboration analysis for distributed datasets

Differentially Private Federated Bayesian Optimization with Distributed Exploration

Contact Info

Product

Resources

About