A bounded-noise mechanism for differential privacy

Dagan, Yuval; Kur, Gil

doi:10.48550/arxiv.2012.03817

Cited by 3 publications

(4 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Another exciting question is to search for distributions such that (8) is tight, where [DK20, GKM20] can serve as good hints. Since [DK20] uses bounded noise, a CLT compatible with bounded log-concave noise is also desirable.…”

Section: Discussionmentioning

confidence: 99%

A Central Limit Theorem for Differentially Private Query Answering

Dong,

Su,

Zhang

2021

Preprint

View full text Add to dashboard Cite

Perhaps the single most important use case for differential privacy is to privately answer numerical queries, which is usually achieved by adding noise to the answer vector. The central question, therefore, is to understand which noise distribution optimizes the privacy-accuracy trade-off, especially when the dimension of the answer vector is high. Accordingly, extensive literature has been dedicated to the question and the upper and lower bounds have been matched up to constant factors [BUV18, SU17]. In this paper, we take a novel approach to address this important optimality question. We first demonstrate an intriguing central limit theorem phenomenon in the high-dimensional regime. More precisely, we prove that a mechanism is approximately Gaussian Differentially Private [DRS21] if the added noise satisfies certain conditions. In particular, densities proportional to e − x α p , where x p is the standard p -norm, satisfies the conditions. Taking this perspective, we make use of the Cramer-Rao inequality and show an "uncertainty principle"-style result: the product of the privacy parameter and the 2 -loss of the mechanism is lower bounded by the dimension. Furthermore, the Gaussian mechanism achieves the constant-sharp optimal privacy-accuracy trade-off among all such noises. Our findings are corroborated by numerical experiments.

show abstract

Section: Discussionmentioning

confidence: 99%

A Central Limit Theorem for Differentially Private Query Answering

Dong,

Su,

Zhang

2021

Preprint

View full text Add to dashboard Cite

show abstract

“…• Bounded Laplace Noise Mechanism: [14] This algorithm adds independent noise in each of the k coordinates, drawn from the distribution µ DE,R that is supported in…”

Section: B Perturbation Dp Mechanismsmentioning

confidence: 99%

Intrusion Detection Based on Privacy-Preserving Federated Learning for the Industrial IoT

Ruzafa-Alcazar

Fernandez-Saura

Marmol-Campos

et al. 2023

IEEE Trans. Ind. Inf.

View full text Add to dashboard Cite

Federated Learning (FL) has attracted significant interest given its prominent advantages and applicability in many scenarios. However, it has been demonstrated that sharing updated gradients/weights during the training process can lead to privacy concerns. In the context of the Internet of Things (IoT), this can be exacerbated due to Intrusion Detection Systems (IDS), which are intended to detect security attacks by analyzing the devices' network traffic. Our work provides a comprehensive evaluation of Differential Privacy (DP) techniques, which are applied during the training of an FL-enabled IDS for Industrial IoT (IIoT). Unlike previous approaches, we deal with noniid data over the recent ToN IoT dataset and compare the accuracy obtained considering different privacy requirements and aggregation functions, namely FedAvg and the recently proposed Fed+. According to our evaluation, the use of Fed+ in our setting provides similar results even when noise is included in the federated training process.

show abstract

“…The multi-label voting problem can also be formulated as simultaneously answering multiple counting queries which has been studied extensively in the theoretical literature, e.g., [15]. These have yet to show promise empirically in settings such as multi-label voting.…”

Section: Prior Work On Multi-label Classificationmentioning

confidence: 99%

“…Pascal VOC 2012 contains 11, 540 images that are split into 5, 717 images for training and 5, 823 images for validation [22]. There are 20 classes of object labels (with their index in parentheses) -aeroplane (1), bicycle (2), bird (3), boat (4), bottle (5), bus (6), car (7), cat (8), chair (9), cow (10), dining table (11), dog (12), horse (13), motorbike (14), person (15), potted plant (16), sheep (17), sofa (18), train (19), and tv monitor (20). We use a ResNet-50 model [25] that was pre-trained on ImageNet [16].…”

Section: Datasets and Model Architecturesmentioning

confidence: 99%

Private Multi-Winner Voting for Machine Learning

Dziedzic

Choquette-Choo

Dullerud

et al. 2023

PoPETs

View full text Add to dashboard Cite

Private multi-winner voting is the task of revealing k-hot binary vectors satisfying a bounded differential privacy (DP) guarantee. This task has been understudied in machine learning literature despite its prevalence in many domains such as healthcare. We propose three new DP multi-winner mechanisms: Binary, Tau, and Powerset voting. Binary voting operates independently per label through composition. Tau voting bounds votes optimally in their L2 norm for tight data-independent guarantees. Powerset voting operates over the entire binary vector by viewing the possible outcomes as a power set. Our theoretical and empirical analysis shows that Binary voting can be a competitive mechanism on many tasks unless there are strong correlations between labels, in which case Powerset voting outperforms it. We use our mechanisms to enable privacy-preserving multi-label learning in the central setting by extending the canonical single-label technique: PATE. We find that our techniques outperform current state-of-the-art approaches on large, real-world healthcare data and standard multi-label benchmarks. We further enable multi-label confidential and private collaborative (CaPC) learning and show that model performance can be significantly improved in the multi-site setting.

show abstract

A bounded-noise mechanism for differential privacy

Cited by 3 publications

References 23 publications

A Central Limit Theorem for Differentially Private Query Answering

A Central Limit Theorem for Differentially Private Query Answering

Intrusion Detection Based on Privacy-Preserving Federated Learning for the Industrial IoT

Private Multi-Winner Voting for Machine Learning

Contact Info

Product

Resources

About