New method improves privacy risk evaluation of machine learning models.
problem Machine learning models can be vulnerable to membership inference attacks.
method Proposed new inference attack method based on prediction entropy, and introduced privacy risk score metric.
result Existing defense approaches are not as effective as previously reported.
New method calibrates noise for attack risk, improving ML model accuracy.
problem Improving accuracy of privacy-preserving ML models while maintaining privacy.
method Directly calibrates noise scale to a desired attack risk level, bypassing the standard ε-calibration. result Significantly decreases noise scale, leading to increased utility at the same risk level.
The calibration of noise for a privacy-preserving mechanism depends on the sensitivity of the query and the prescribed privacy level. A data steward must make the non-trivial choice of a privacy level that balances the requirements of users and the monetary constraints of the business entity. We analyse roles of the so…
New privacy framework tailored to specific data distributions.
problem Protecting individual data points in decision-making processes.
method Introducing tangent differential privacy, a new form of differential privacy.
result Entropic regularization guarantees tangent differential privacy under general conditions.
This paper introduces a novel approach to measuring privacy risks in deep computer vision models based on intermediate outputs.
problem The exposure of intermediate results in hidden layers of deep computer vision models poses significant privacy concerns.
method The approach leverages Degrees of Freedom (DoF) to evaluate the amount of information retained in each layer and combines this with the rank of the Jacobian matrix to assess sensitivity to input variations.
result The proposed framework provides deeper insights into privacy risks associated with intermediate representations without requiring adversarial attack simulations.
Privacy affects how much data is needed for CVaR optimization.
problem Privacy constraints impact the effective sample size for CVaR optimization.
method Analyzes the privacy-relevant sample size and decomposes CVaR excess risk.
result The effective private tail sample size is εnτ, affecting CVaR learning rates.
Unified framework analyzes privacy risks from gradients in distributed learning.
problem Analyzing inference privacy risks from gradients in machine learning.
method Unified game-based framework for various attacks, including attribute, property, distributional, and user disclosures.
result Demonstrates inefficacy of data aggregation for privacy against inference attacks.
Fine-tuning LLMs on privacy-sensitive data introduces privacy risk, and synthetic data audits can quantify this risk.
problem Fine-tuning LLMs on privacy-sensitive data introduces privacy risk.
method Generate synthetic canaries via high-temperature sampling from LLMs.
result Synthetic canaries are high-influence outliers that ensure strong audits.
Paper improves privacy in SGD with low noise, achieving optimal risk rates.
problem Privacy-preserving machine learning with good performance.
method Differentially private SGD with low-noise analysis.
result Achieves optimal excess risk rates for non-smooth losses.
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.
Score attack method provides a lower bound on privacy-constrained minimax risk.
problem Characterizing the optimality of privacy-constrained statistical models.
method Score attack based on tracing attack concept.
result Optimally lower bounds the minimax risk of estimating unknown model parameters.
Fairness in machine learning increases privacy risks, especially for underrepresented groups.
problem Privacy risks in fair machine learning models, particularly for underrepresented groups.
method Membership inference attacks to measure information leakage and analyze fairness vs. privacy trade-offs.
result Achieving fairness in machine learning models increases privacy risks, especially for underrepresented groups.
This paper addresses privacy concerns in ratio statistics using differential privacy.
problem Privacy concerns in ratio statistics across machine learning areas.
method Develops a simple algorithm for differentially private ratio statistics, proving consistency and constructing confidence intervals.
result A simple algorithm can provide excellent privacy, sample accuracy, and bias properties in ratio statistics.
Paper introduces LR to generate synthetic data with privacy protection.
problem Privacy concerns limit the use of sensitive datasets.
method Local Resampler (LR) using k-nearest neighbors algorithm.
result LR effectively mitigates outlier-driven disclosure risks.
Improved privacy and utility in machine learning with adaptive differential privacy.
problem Enhancing privacy in machine learning models while maintaining utility.
method Adaptive differentially private (ADP) learning method that optimally adapts noise to stepsize.
result ADP method significantly improves utility compared to standard differentially private methods.
FSL-BDP models time-to-default without centralizing data, improving privacy mechanisms in federated settings.
problem Traditional credit risk models ignore default timing and violate data-protection rules.
method Federated Survival Learning with Bayesian Differential Privacy (FSL-BDP).
result FSL-BDP improves privacy mechanisms in federated settings, outperforming classical DP in most clients.
BUDS balances privacy and utility by shuffling data, achieving strong privacy with minimal loss.
problem Balancing privacy and utility in crowd-sourced statistical databases.
method One-hot encoding, iterative shuffling, loss estimation, risk minimization.
result Achieves ε=0.02 for privacy, maintaining a privacy bound of ε=ln[t/((n1−1)S)]. Offline RL algorithms protect privacy while learning from sensitive data.
problem Learning sensitive data in financial, legal, and healthcare applications.
method Differential privacy guarantees for offline RL algorithms.
result Provable prevention of privacy risks with strong learning bounds.
Enhanced stability improves privacy in machine learning.
problem Improving privacy in machine learning training while maintaining accuracy.
method Study of stability in private empirical risk minimization, focusing on strongly-convex loss functions and uniform stability.
result An algorithm with uniform stability of β implies a bound of O(√β) on the scale of noise required for differential privacy.
Paper studies optimal federated learning for nonparametric regression with privacy constraints.
problem Federated learning for nonparametric regression with heterogeneous differential privacy constraints.
method Proposes distributed privacy-preserving estimators and investigates their risk properties.
result Establishes matching minimax lower bounds for global and pointwise estimation.
Differential Privacy (DP) provides strong guarantees on the risk of compromising a user's data in statistical learning applications, though these strong protections make learning challenging and may be too stringent for some use cases. To address this, we propose element level differential privacy, which extends differ…
NAPP-ERM improves ERM with differential privacy guarantees by iteratively achieving target regularization and delivering strong convexity.
problem Over-regularization in privacy-preserving ERM approaches.
method Noise-Augmented Privacy-Preserving Empirical Risk Minimization (NAPP-ERM) with a dual-purpose l2 regularizer and privacy budget retrieval strategy.
result Mitigates over-regularization and achieves strong convexity through a single regularizer.
Unified bounds for DP risks reduce noise and improve accuracy.
problem Difficult interpretation and calibration of DP mechanisms.
method Hypothesis-testing interpretation of DP (f-DP) and unified bounds. result Unified bounds are tighter and tunable for specific risks.
New method assesses individual training points' privacy risk without retraining.
problem Privacy vulnerability of individual training points in membership inference attacks.
method Derives a closed-form decomposition of individual black-box MIA vulnerability, extending to deep networks.
result Proposes a surrogate score operating on last-layer representations that requires only a single trained model.
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.
A new DP algorithm for weighted ERM protects sensitive data in predictive models.
problem Protecting sensitive personal information in predictive models trained via ERM.
method Proposes the first differentially private algorithm for weighted ERM with formal privacy guarantees.
result Demonstrates strong DP guarantees while maintaining robust performance in real-world data.
This paper analyzes privacy risks in neural network pruning and proposes a defense mechanism.
problem Privacy risks in neural network pruning due to membership inference attacks.
method Investigates the impact of pruning on prediction divergence and proposes a self-attention membership inference attack.
result Proposed defense mechanism mitigates privacy risks while maintaining sparsity and accuracy.
Introduces six levels of privacy for financial synthetic data.
problem Privacy risks in financial synthetic data.
method Defines six levels of privacy protection for synthetic data.
result Provides a framework to assess and improve privacy in synthetic data.
This paper introduces DPI, a new metric to assess data-copying risk in tabular data.
problem Measuring privacy risk of data-copying in tabular generative models.
method Proposes Data Plagiarism Index (DPI) for evaluating data-copying risk.
result DPI identifies data-copying threats to tabular data models, highlighting privacy and fairness issues.
We study the problem of data release with privacy, where data is made available with privacy guarantees while keeping the usability of the data as high as possible --- this is important in health-care and other domains with sensitive data. In particular, we propose a method of masking the private data with privacy guar…
The study tightens risk bounds for mixtures of experts using local differential privacy.
problem Improving risk bounds for mixtures of experts.
method Imposing local differential privacy (LDP) on the gating mechanism of mixtures of experts.
result Theoretical bounds exhibit logarithmic dependence on the number of experts and tighter than existing bounds.
Machine unlearning can compromise privacy, study shows.
problem Machine unlearning may leave data imprints in ML models, risking privacy.
method Proposed a membership inference attack to detect leakage.
result Machine unlearning can lead to unintended privacy risks.
Synthetic data improves credit scoring models' performance without compromising borrower privacy.
problem Scarcity of real data for credit scoring models due to privacy concerns.
method Privacy-preserving training with synthetic data.
result Credit scoring models trained with synthetic data show a reduction of 3% in AUC and 6% in KS compared to real data models.
Membership inference attacks seek to infer the membership of individual training instances of a privately trained model. This paper presents a membership privacy analysis and evaluation system, called MPLens, with three unique contributions. First, through MPLens, we demonstrate how membership inference attack methods …
Optimizes differentially private kernel learning with random projection.
problem Privacy-preserving learning algorithms with optimal performance.
method Differentially private kernel ERM algorithm based on random projection in reproducing kernel Hilbert space.
result Achieves minimax-optimal excess risk rates for various loss functions.
Study creates a multimodal learning framework for CVD risk prediction.
problem Predicting cardiovascular disease risk in diverse populations.
method Combines cross modal transformers, graph neural networks, and causal representation learning.
result Model predicts personalized CVD risk with causal invariance across subpopulations.
Large capacity machine learning (ML) models are prone to membership inference attacks (MIAs), which aim to infer whether the target sample is a member of the target model's training dataset. The serious privacy concerns due to the membership inference have motivated multiple defenses against MIAs, e.g., differential pr…
We study statistical risk minimization problems under a privacy model in which the data is kept confidential even from the learner. In this local privacy framework, we establish sharp upper and lower bounds on the convergence rates of statistical estimation procedures. As a consequence, we exhibit a precise tradeoff be…
The study introduces a holdout-based framework to assess synthetic data fidelity and privacy.
problem Evaluating the quality and privacy of synthetic data solutions for mixed-type tabular data.
method Holdout-based empirical assessment framework measuring fidelity and privacy risk.
result Synthetic data samples are as close to the training as to the holdout data, indicating generalization and independence from individual records.
This paper studies fairness and privacy in federated learning, proposing algorithms to balance both.
problem Joint impact of differential privacy and fairness in federated classification.
method Proposes FDP-Fair and CDP-Fair algorithms for demographic disparity constrained classification under federated differential privacy.
result Established theoretical guarantees on privacy, fairness, and excess risk control.
Synth-MIA assesses privacy leakage in synthetic tabular data models.
problem Challenges in evaluating privacy leakage in synthetic tabular data.
method Unified threat framework deploying multiple attacks.
result Higher synthetic data quality correlates with greater privacy leakage.
New algorithm optimizes privacy and utility in multi-task learning with skewed data.
problem Privacy constraints in multi-task learning with uneven data distribution.
method Adaptive reweighting of privacy budget allocation among tasks.
result Significant improvement in utility with state-of-the-art performance on benchmarks.
Differential privacy is concerned about the prediction quality while measuring the privacy impact on individuals whose information is contained in the data. We consider differentially private risk minimization problems with regularizers that induce structured sparsity. These regularizers are known to be convex but they…
The collection and analysis of user data drives improvements in the app and web ecosystems, but comes with risks to privacy. This paper examines discrete distribution estimation under local privacy, a setting wherein service providers can learn the distribution of a categorical statistic of interest without collecting …
Privacy and transparency are two key foundations of trustworthy machine learning. Model explanations offer insights into a model's decisions on input data, whereas privacy is primarily concerned with protecting information about the training data. We analyze connections between model explanations and the leakage of sen…
Develops new instance-optimality concepts in differential privacy.
problem Improving privacy guarantees in statistical estimation.
method Introduces local minimax risk and unbiased mechanisms, and develops inverse sensitivity mechanisms.
result Inverse sensitivity mechanisms are nearly instance optimal for a wide range of functions.
Framework for AI customer support that protects privacy and reduces costs.
problem Privacy risks and compliance challenges in AI customer support.
method Zero-shot learning with large language models, real-time data anonymization, retrieval-augmented generation, robust post-processing.
result Reduces privacy risks and compliance costs while maintaining accuracy.
Paper evaluates and mitigates privacy risks in deep learning models.
problem Quantifying and defending against privacy attacks in deep learning.
method Quantitative evaluation of trade-offs, reformulating attacks, and proposing a novel SPN.
result Model accuracy improved by 5-20% while maintaining data privacy.