Algorithm selects public datasets for private machine learning.
problem Choosing the most suitable public dataset for private machine learning.
method Measures gradient subspace distance between public and private datasets.
result Excess risk scales with the subspace distance between gradients.
Paper evaluates and improves private feature selection methods.
problem Feature selection in high-dimensional datasets with privacy constraints.
method Correlations-based order statistic privatized for feature selection.
result Our method significantly outperforms established baseline for private feature selection.
Many applications of machine learning, such as human health research, involve processing private or sensitive information. Privacy concerns may impose significant hurdles to collaboration in scenarios where there are multiple sites holding data and the goal is to estimate properties jointly across all datasets. Differe…
Improved image generation with private data using perceptual features.
problem Difficulty in training generative models with differential privacy.
method Use pre-trained perceptual features to learn private data distribution.
result Generative models can generate high-quality images with low privacy budget (ϵ≈2). A key factor in developing high performing machine learning models is the availability of sufficiently large datasets. This work is motivated by applications arising in Software as a Service (SaaS) companies where there exist numerous similar yet disjoint datasets from multiple client companies. To overcome the challen…
First private Bayesian optimization algorithm with provable performance.
problem Privacy-preserving Bayesian optimization in outsourced settings.
method Random projection-based transformation to preserve privacy and distances.
result Regret bound similar to non-private GP-UCB established for PO-GP-UCB.
PriMORL trains private RL policies on offline data.
problem Private reinforcement learning on offline data.
method PriMORL learns DP models of the environment and optimizes a policy on the penalized private model.
result PriMORL enables training of private RL agents on complex tasks.
First DP algorithm for Wasserstein barycenters on private data.
problem Computing Wasserstein barycenters on private datasets.
method Differentially private algorithms for Wasserstein barycenters.
result High-quality private barycenters with strong accuracy-privacy tradeoffs.
Differentially private learning avoids iterative optimization in parameter space.
problem Differentially private training of neural networks
method Hypernetworks trained on public datasets
result Significantly reduces noise in parameter space
We envision AI marketplaces to be platforms where consumers, with very less data for a target task, can obtain a relevant model by accessing many private data sources with vast number of data samples. One of the key challenges is to construct a training dataset that matches a target task without compromising on privacy…
DPLTM uses private winners to improve privacy and utility.
problem Balancing privacy and utility in machine learning models.
method DPLTM applies differential privacy to the lottery ticket method.
result DPLTM converges faster and uses less privacy budget.
We study the problem of estimating high dimensional models with underlying sparse structures while preserving the privacy of each training example. We develop a differentially private high-dimensional sparse learning framework using the idea of knowledge transfer. More specifically, we propose to distill the knowledge …
Private sketches protect linear regression data privacy.
problem Protecting sensitive information in linear regression.
method Release private sketches of datasets, compute approximate solutions.
result Private sketches maintain good approximation guarantees to the original problem.
Paper introduces algorithms for private decision tree learning.
problem Private decision tree learning in distributed settings.
method Proposes DP-TopDown, NoisyCounts, and LocalRNM.
result First utility guarantees for private decision tree learning.
FPFL mitigates unfairness in private federated learning.
problem Differential privacy degrades model performance on under-represented groups.
method Extends modified method of differential multipliers to private federated learning.
result FPFL reduces unfairness in trained models on private federated learning.
New algorithms improve privacy and utility of large language models.
problem Privacy-preserving fine-tuning of large language models.
method Meta-framework for differentially private fine-tuning, inspired by recent success in fine-tuning.
result Private fine-tuned models achieve utility close to non-private models, with improved privacy and efficiency.
PASS protects private attributes by stochastically substituting data.
problem Protecting private attributes in ML services while maintaining data utility.
method PASS uses stochastic data substitution with a novel loss function derived from information theory.
result PASS effectively protects private attributes across various datasets.
Private PGB boosts synthetic data quality using GANs and privacy techniques.
problem Differentially private GANs struggle with convergence and poor output quality.
method Combines reweighted samples from GAN training using Private Multiplicative Weights method.
result Improves synthetic data quality across various datasets and tasks.
Study shows privacy and utility trade-offs in synthetic data models, impacting fairness and real-world performance.
problem Understanding the impact of differential privacy on fairness and model performance in synthetic data.
method Systematic analysis of differentially private synthetic datasets on classification models, measuring utility and bias using fairness metrics.
result More privacy does not necessarily mean more bias, but it can affect model performance when deployed on real data.
Study privacy and robustness in learning halfspaces, proving hard trade-offs.
problem Balancing privacy and robustness in learning halfspaces.
method Proves nearly tight bounds on sample complexity for robust private learning of halfspaces.
result Robust and private learning is harder than robust or private learning alone.
Motivation: Human genomic datasets often contain sensitive information that limits use and sharing of the data. In particular, simple anonymisation strategies fail to provide sufficient level of protection for genomic data, because the data are inherently identifiable. Differentially private machine learning can help b…
DP-CGAN generates private synthetic data and labels.
problem Preserving privacy in synthetic data generation.
method Differentially private conditional GAN (DP-CGAN) with clipping and perturbation.
result DP-CGAN generates visually and empirically promising results on MNIST with low privacy cost.
Efficiently estimates private least squares with linear error growth.
problem Private estimation of ordinary least squares with bounded residuals and leverage.
method Scaled noise added to a stable nonprivate estimator of the regression vector.
result Near-optimal accuracy guarantee with linear error growth in dimension.
New KD-tree based method for private synthetic data generation.
problem Creating private synthetic data that accurately represents real data.
method KD-trees combined with noise perturbation for differentially private synthetic data generation.
result Our data-dependent approach improves utility over prior work and scales well.
In this paper, we study the problem of precision matrix estimation when the dataset contains sensitive information. In the differential privacy framework, we develop a differentially private ridge estimator by perturbing the sample covariance matrix. Then we develop a differentially private graphical lasso estimator by…
Asynchronous algorithms reduce privacy costs in distributed machine learning.
problem Privacy concerns in training machine learning models on scattered private data.
method Differentially-private asynchronous algorithms for collaborative training.
result Cost of privacy is inversely proportional to dataset size and privacy budgets.
Due to massive amounts of data distributed across multiple locations, distributed machine learning has attracted a lot of research interests. Alternating Direction Method of Multipliers (ADMM) is a powerful method of designing distributed machine learning algorithm, whereby each agent computes over local datasets and e…
Paper develops private synthetic data for sensitive data analysis.
problem Analyzing sensitive data while preserving privacy.
method Adapting a private density estimation algorithm for synthetic data generation.
result Better computational guarantees for synthetic data generation.
This paper sharpens privacy guarantees for high-dimensional PCA under differential privacy.
problem Understanding the exact privacy loss in high-dimensional PCA with differential privacy.
method Analyzes the exponential mechanism in a model-free setting for high-dimensional PCA.
result Sharp utility and privacy characterizations in high dimensions show the difficulty of detecting a target individual's presence.
Catastrophic forgetting can be a significant problem for institutions that must delete historic data for privacy reasons. For example, hospitals might not be able to retain patient data permanently. But neural networks trained on recent data alone will tend to forget lessons learned on old data. We present a differenti…
Privacy-preserving synthetic data from EHRs for learning and inference.
problem Sharing sensitive EHR data while maintaining patient privacy.
method Differentially private normalizing flows for density estimation and variational inference.
result Privacy-preserving synthetic data can yield good utility at a reasonable privacy cost.
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.
Blind source separation algorithms such as independent component analysis (ICA) are widely used in the analysis of neuroimaging data. In order to leverage larger sample sizes, different data holders/sites may wish to collaboratively learn feature representations. However, such datasets are often privacy-sensitive, prec…
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.
Differentially private ensemble classifiers adapt to data streams while protecting privacy.
problem Adapting to evolving data characteristics while protecting private information.
method Unbounded ensemble updates, model agnostic approach.
result Outperforms competitors on various privacy, drift, and distribution settings.
Few-shot DP image classification models need more data as privacy increases.
problem Few-shot DP image classification challenges in personalization and federated learning.
method Exhaustive experiments on various parameters affecting few-shot DP accuracy and vulnerability.
result Increasing shots per class is necessary for DP accuracy as privacy increases.
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.
PILLAR improves SP learning with less private data.
problem Efficiently learning with semi-private data under privacy constraints.
method Uses pre-trained public data features to reduce private data requirements.
result Significantly lower private labelled sample complexity achieved.
Publicly pretraining models on Web data may undermine differential privacy.
problem The use of large Web-scraped datasets in differential privacy models.
method Critical review of leveraging pretrained models on public datasets for differential privacy.
result Publicizing pretrained models as 'private' could harm trust and generalize poorly.
Paper improves privacy-accuracy balance in federated learning.
problem Privacy-accuracy tradeoffs in federated learning.
method Personalized federated learning with joint differential privacy.
result Coordination of local and centralized learning improves accuracy while maintaining privacy.
Differentially private conformal prediction improves statistical efficiency.
problem Quantifying uncertainty in private data analysis.
method Introducing differential conformal prediction and developing Differentially Private Conformal Prediction (DPCP).
result DPCP produces tighter prediction sets than existing private split conformal approaches.
Study differentially private methods for learning Hawkes processes.
problem Lack of thorough analysis on sample complexity for learning Hawkes processes parameters and releasing differentially private versions.
method Developed non-private and differentially private estimators for Hawkes processes parameters.
result Obtained sample complexity results for both private and non-private settings.
Paper introduces DP TDA for near-optimal private persistence diagrams.
problem Challenges in privatizing topological data analysis.
method Sensitivity analysis of persistence diagrams, use of exponential mechanism.
result Proposes near-optimal privacy mechanism for TDA.
Private training and synthetic data generation using DP clustering.
problem Protecting sensitive data in deep neural networks training.
method Approximate input dataset with privately generated synthetic dataset using DP clustering.
result Simple two-layer neural network achieves SOTA classification accuracy on standard benchmark datasets.
Discond-VAE separates continuous and discrete factors in data.
problem Separating shared and class-specific variations in real-world data.
method Introduces private and public latent variables to represent continuous and discrete factors, respectively.
result Discond-VAE successfully disentangles class-dependent continuous factors from discrete factors.
EIGAN learns private representations without centralized data, outperforming state-of-the-art.
problem Private representation learning with multiple ally and adversary attributes.
method Exclusion-Inclusion Generative Adversarial Network (EIGAN) and Distributed EIGAN (D-EIGAN).
result EIGAN and D-EIGAN outperform state-of-the-art methods in accuracy and scalability.
Optimizes sparse fine-tuning for privacy in neural networks.
problem Performance gap between DP-SGD and non-private fine-tuning.
method Optimization-based approach using private gradient information for selecting trainable weights.
result Our selection method leads to better prediction accuracy compared to existing approaches.
Survival function estimation is used in many disciplines, but it is most common in medical analytics in the form of the Kaplan-Meier estimator. Sensitive data (patient records) is used in the estimation without any explicit control on the information leakage, which is a significant privacy concern. We propose a first d…