This work optimizes mean estimation under varying privacy constraints.
problem Mean estimation with heterogeneous privacy constraints.
method Proposes an algorithm for mean estimation under different privacy levels for users.
result Shows a saturation phenomenon in performance as privacy levels are relaxed.
Paper relaxes differential privacy for correlated features, improving privacy-utility trade-off.
problem Standard differential privacy ignores feature correlation, leading to suboptimal privacy-utility balance.
method Introduces CorrDP framework that accounts for feature correlation, using total variation distance for quantification.
result CorrDP algorithms outperform standard DP in synthetic and real-world datasets with insensitive features.
Differential privacy has seen remarkable success as a rigorous and practical formalization of data privacy in the past decade. This privacy definition and its divergence based relaxations, however, have several acknowledged weaknesses, either in handling composition of private algorithms or in analyzing important primi…
Parameter-transfer is a well-known and versatile approach for meta-learning, with applications including few-shot learning, federated learning, and reinforcement learning. However, parameter-transfer algorithms often require sharing models that have been trained on the samples from specific tasks, thus leaving the task…
Novel privacy model for decentralized data analysis.
problem Achieving good privacy-utility trade-off in federated learning.
method Introducing network Differential Privacy (network DP) for decentralized algorithms.
result Privacy-utility trade-offs of network DP algorithms significantly improve upon LDP and trusted curator model.
We consider the problem of reinforcing federated learning with formal privacy guarantees. We propose to employ Bayesian differential privacy, a relaxation of differential privacy for similarly distributed data, to provide sharper privacy loss bounds. We adapt the Bayesian privacy accounting method to the federated sett…
The Sampled Gaussian Mechanism (SGM)---a composition of subsampling and the additive Gaussian noise---has been successfully used in a number of machine learning applications. The mechanism's unexpected power is derived from privacy amplification by sampling where the privacy cost of a single evaluation diminishes quadr…
Large data collections required for the training of neural networks often contain sensitive information such as the medical histories of patients, and the privacy of the training data must be preserved. In this paper, we introduce a dropout technique that provides an elegant Bayesian interpretation to dropout, and show…
This paper proposes a new differential privacy definition using Rao distance.
problem Improving differential privacy definitions for better sequential composition.
method Using Rao distance instead of divergences of densities to define privacy.
result Proposed definition shares interpretation with previous definitions but improves sequential composition.
Differential privacy, a notion of algorithmic stability, is a gold standard for measuring the additional risk an algorithm's output poses to the privacy of a single record in the dataset. Differential privacy is defined as the distance between the output distribution of an algorithm on neighboring datasets that differ …
Differential privacy is a de facto standard in data privacy, with applications in the public and private sectors. A way to explain differential privacy, which is particularly appealing to statistician and social scientists is by means of its statistical hypothesis testing interpretation. Informally, one cannot effectiv…
New divergences help audit DP in high dimensions.
problem Challenges in auditing DP in high-dimensional data.
method Propose kernel Rényi divergence and its regularized version for auditing.
result Regularized kernel Rényi divergence can be estimated from samples in high dimensions.
Popular approaches to differential privacy, such as the Laplace and exponential mechanisms, calibrate randomised smoothing through global sensitivity of the target non-private function. Bounding such sensitivity is often a prohibitively complex analytic calculation. As an alternative, we propose a straightforward sampl…
The paper develops methods to create private synthetic spatial point patterns.
problem Generating private synthetic spatial point patterns.
method Developed differentially private Poisson and Cox point synthesizers.
result The synthesizers effectively maintain privacy and utility of synthetic data.
Deep learning models are often trained on datasets that contain sensitive information such as individuals' shopping transactions, personal contacts, and medical records. An increasingly important line of work therefore has sought to train neural networks subject to privacy constraints that are specified by differential…
Paper improves differential privacy analysis for machine learning.
problem Quantifying privacy leakage in noisy gradient descent.
method Shifted interpolation in f-differential privacy. result First exact privacy analysis for strongly convex optimization.
Paper relaxes SGD privacy and generalization guarantees for non-smooth convex losses.
problem Privacy and generalization in SGD for non-smooth convex losses.
method Relaxes Lipschitz and strong smoothness assumptions to Hölder smoothness, proving (ε,δ)-DP and optimal excess risk. result Noisy SGD with α-Hölder smooth losses achieves optimal excess risk with linear gradient complexity for α≥1/2. Deep neural networks with their large number of parameters are highly flexible learning systems. The high flexibility in such networks brings with some serious problems such as overfitting, and regularization is used to address this problem. A currently popular and effective regularization technique for controlling the…
New framework provides privacy guarantees for practical federated learning.
problem Inadequate privacy guarantees for federated learning due to restrictive assumptions.
method Fed-α-NormEC, integrating multiple local updates, partial client participation, and standard assumptions. result Provably convergent and differentially private federated learning framework.
Federated learning approach for binary matrix factorization.
problem Efficiently factorizing binary data distributed across stakeholders while maintaining privacy.
method Proximal optimization for federated learning of relaxed binary matrix factorization.
result Federated algorithm outperforms state-of-the-art methods in quality and efficacy.
Decor protects decentralized learning models from curious users.
problem Privacy violation in decentralized learning.
method Decor uses correlated Gaussian noises to protect local models in decentralized SGD with differential privacy guarantees.
result Decor matches central DP optimal privacy-utility trade-off for arbitrary connected graphs.
Iteratively reweighted least squares (IRLS) is a widely-used method in machine learning to estimate the parameters in the generalised linear models. In particular, IRLS for L1 minimisation under the linear model provides a closed-form solution in each step, which is a simple multiplication between the inverse of the we…
New credit attribution methods for machine learning models using relaxed stability guarantees.
problem Ensuring proper attribution in generative models trained on existing works.
method Proposed new definitions of stability that allow for non-stable processing of a subset of datapoints with permission.
result Extended well-studied stability notions and provided a comprehensive characterization of learnability.
We compare the sample complexity of private learning [Kasiviswanathan et al. 2008] and sanitization~[Blum et al. 2008] under pure ε-differential privacy [Dwork et al. TCC 2006] and approximate (ε,δ)-differential privacy [Dwork et al. Eurocrypt 2006]. We show that the sample complexity of these tasks under approxima…
Paper optimizes private ReLU regression with relaxed assumptions.
problem Private ReLU regression with relaxed assumptions.
method One-pass mini-batch Generalized Linear Model Perceptron algorithm (DP-MBGLMtron).
result Achieves optimal utility bound up to logarithmic factors.
Improved privacy in RL with near-optimal regret bounds.
problem Privacy-preserving reinforcement learning in personalized decision-making systems.
method Differentially private algorithm based on LSVI-UCB++ with privacy-preserving techniques.
result Achieved a near-optimal regret bound of O(d * sqrt(H^3 * K) + H^(15/4) * d^(7/6) * K^(1/2) / ε).
New bounds on private mean estimation for heavy-tailed distributions.
problem Estimating the mean of heavy-tailed distributions under differential privacy constraints.
method Upper and lower bounds on sample complexity for differentially private mean estimation.
result Qualitatively different sample complexity compared to non-private estimation, with a factor of O(d) larger for multivariate cases. Polynomial-time private algorithm for robust estimation of mean and covariance in the presence of outliers.
problem Estimating mean and covariance in the presence of adversarial outliers.
method Stabilizing convex relaxations using a new estimate-dependent noise injection mechanism.
result First efficient private robust estimation algorithm for covariance without condition-number assumptions.
This paper enhances privacy in statistical model checking of cyber-physical systems.
problem Privacy concerns in consumer-level applications due to statistical model checking.
method Proposes expected differential privacy and a new exponential mechanism for sequential algorithms.
result Demonstrates a novel mechanism to preserve privacy in statistical model checking.
Paper improves differential privacy SGD by considering data heterogeneity.
problem Improving differential privacy in machine learning with varying data contributions.
method Introducing influence function to quantify data contributions, designing PIDP-SGD algorithm.
result PIDP-SGD significantly improves machine learning model performance.
Private mean estimation with multiple samples requires a certain number of people to maintain privacy.
problem Private mean estimation with person-level differential privacy for multiple samples.
method The approach involves estimating the mean up to a distance α in ℓ_2-norm under ε-differential privacy, using algorithms based on the clip-and-noise framework and new analyses.
result The necessary and sufficient number of people to estimate the mean up to distance α in ℓ_2-norm is given by a specific formula.
Private learning is hard when data is long-tailed.
problem Achieving both privacy and fairness in machine learning with long-tailed data.
method Theoretical analysis and experimental validation on various datasets and algorithms.
result Relaxing overall accuracy can lead to good fairness even with strict privacy requirements.
Data privacy is an important concern in learning, when datasets contain sensitive information about individuals. This paper considers consensus-based distributed optimization under data privacy constraints. Consensus-based optimization consists of a set of computational nodes arranged in a graph, each having a local ob…
Paper uses statistical depth to create DP estimators for regression.
problem Creating differentially private estimators in high dimensions.
method Uses halfspace and regression depth to analyze maximum influence and construct DP estimators.
result New DP estimators for location and regression show favorable performance.
New regularizer for machine learning using private data.
problem Machine learning with private data.
method Distributionally-robust optimization with locally-differentially-private datasets.
result New regularizer for training linear regression models.
Optimal Gaussian noise mechanisms achieve nearly optimal error in unbiased mean estimation.
problem Efficiently estimating the mean of high-dimensional data while preserving privacy.
method Differential privacy mechanisms with Gaussian noise, focusing on optimal covariance.
result Gaussian noise mechanisms achieve nearly optimal error among all private unbiased mean estimation mechanisms.
Through the lens of information-theoretic reductions, we examine a reductions approach to fair optimization and learning where a black-box optimizer is used to learn a fair model for classification or regression. Quantifying the complexity, both statistically and computationally, of making such models satisfy the rigor…
Paper proposes distributed optimization for federated learning with theoretical guarantees.
problem Privacy-preserving cross-organizational data collaboration in machine learning.
method Augmented Lagrangian technique for diverse communication topologies, termination criteria, and parameter update mechanisms.
result The proposed framework recovers classical optimization methods and provides strong performance in large-scale federated learning.
We provide a differentially private algorithm for hypothesis selection. Given samples from an unknown probability distribution P and a set of m probability distributions H, the goal is to output, in a ε-differentially private manner, a distribution from H whose total variation di…
Paper improves classification rates for private data.
problem Classifying data with privacy constraints and relaxed assumptions.
method Introduced a novel approach for classification under privacy constraints, relaxing the strong density assumption.
result Achieved minimax optimal convergence rates without strong density assumption.
Study differentially private linear regression with heavy-tailed data.
problem Differentially private ℓ1-norm linear regression with heavy-tailed data. method Exponential mechanism for ℓ2-norm bounded second moment; relaxation to ℓ2-norm bounded θ-th moment; coordinate-wise bounded moments. result Achieved upper bounds for privacy-preserving linear regression under various moment conditions.
Gradient Boosting Decision Trees (GBDTs) have become very successful in recent years, with many awards in machine learning and data mining competitions. There have been several recent studies on how to train GBDTs in the federated learning setting. In this paper, we focus on horizontal federated learning, where data sa…
Improves privacy amplification by shuffling for differential privacy.
problem Enhancing privacy guarantees in systems with anonymous data contributions.
method Theoretical and numerical analysis of Rényi differential privacy parameters and privacy amplification by shuffling.
result First asymptotically optimal analysis of Rényi differential privacy parameters for shuffled outputs.
Proposes a privacy-preserving recommendation system using matrix factorization and differential privacy.
problem Privacy leakage in recommendation systems when anonymizing user data is not sufficient.
method Uses matrix factorization and differential privacy via the Gaussian mechanism.
result Demonstrates excellent utility for privacy-preserving recommendation systems.
Improves privacy guarantees by analyzing randomness in privacy-preserving mechanisms.
problem Balancing user privacy and business constraints in privacy-preserving mechanisms.
method Analyzes explicit and implicit randomness in privacy mechanisms and proposes a probabilistic calibration method.
result Proposes privacy at risk, providing stronger privacy guarantees with quantifiable risks.
The paper analyzes and proposes methods for privately sharing individual privacy losses using per-instance differential privacy.
problem The standard differential privacy framework provides a worst-case bound that may not accurately reflect individual privacy losses.
method The paper analyzes per-instance differential privacy and proposes methods to privately and accurately publish per-instance privacy losses.
result The methods privately and accurately publish per-instance differential privacy losses with minimal additional privacy cost.
This work optimizes mean estimation under varying user privacy demands.
problem Mean estimation with heterogeneous privacy levels.
method Proposes an algorithm that is minimax optimal and has near-linear run-time.
result Privacy requirements of the most stringent users dictate overall error rates.
A new privacy accountant for Gaussian differential privacy measures individual privacy losses.
problem Bounding differential privacy loss for each participant in data analysis.
method Developed a privacy accountant for adaptive compositions of randomised mechanisms using Gaussian differential privacy.
result Provided optimal bounds for the Gaussian mechanism and constructed an approximative individual privacy accountant.