Differentially Private Learning with Kernels

By Prateek Jain, Abhradeep Thakurta. 2013

101 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

One-Shot Federated Learning (2019, background): “For example, it may be that a certain number of devices agree to share their local data, or that there exists some prior publicly available data for the problem at hand [3].”
Differentially Private Empirical Risk Minimization: Efficient Algorithms and Tight Error Bounds (2014, background): “[17], Smith and Thakurta [44] study online optimization, Jain and Thakurta [26] study an interactive model tailored to high-dimensional kernel learning.”
Towards Practical Differentially Private Convex Optimization (2019, neutral): “Many recent works [30], [31], [32] examine the setting of online learning, whereas highdimensional kernel learning is considered in [33]; these settings are quite different from ours, and the results are incomparable.”
Learning with Differential Privacy: Stability, Learnability and the Sufficiency and Necessity of ERM Principle (2015, background): “Related work While a large amount of work has been devoted to finding consistent (and rate optimal) differentially private learning algorithms in various settings (e.g., Chaudhuri et al., 2011; Kifer et al., 2012; Jain and Thakurta, 2013; Bassily et al., 2014), the characterization of privately lear”
Privacy-Preserving Distributed Multi-Task Learning with Asynchronous Updates (2017, methodology): “[9, 10] where an efficient differentially private empirical risk minimization was developed, Jain and Thakurta [26] studied kernel learning under three practical models, and Kusner et al.”
A Semi-Supervised Learning Approach to Differential Privacy (2013, background): “This is done by adding a noise proportional to the sensitivity of f , drawn from a suitably chosen distribution, to the output of f (D ), where D ∈ X .”
(Near) Dimension Independent Risk Bounds for Differentially Private Learning (2014, background): “An exception to the above is the work by Jain & Thakurta (2013) that provides a differentially private method for learning with kernels, however their methods assume a stronger model where a few samples from the unlabeled test data set are also available.”
Uplink Communication Efficient Differentially Private Sparse Optimization With Feature-Wise Distributed Data (2018, methodology): “Beginning with the seminal work (Chaudhuri, Monteleoni, and Sarwate 2011), which considers private ERM training under the formal statistical differential privacy notion (Dwork, Roth, and others 2014), various differentially private optimization algorithms have been developed for training the model w”
Private Empirical Risk Minimization, Revisited (2014, background): “Jain and Thakurta [23] gave dimension-independent expected excess risk bounds for the special case of “generalized linear models” with a strongly convex regularizer, assuming that C = Rp (that is, unconstrained optimization).”
Differentially Private Data Publishing and Analysis: A Survey (2017, methodology): “By choosing different loss functions, ERM can be implemented for certain learning tasks, such as linear regression [69], in which the loss function is defined as maximum likelihood estimation (MLE); logistic regression [68], in which the loss function is defined as logistic loss; and the kernel meth”
Differential Privacy and Machine Learning: a Survey and Review (2014, methodology): “The Test Data-independent Learner (TTDP) mechanism in [26] publishes a private kernel SVM model satisfying (ǫ, δ)-differential privacy as follows.”
When Relaxations Go Bad: "Differentially-Private" Machine Learning (2019, unknown): “…has also been a considerable progress in generalizing privacy-preserving machine learning to more complex scenarios such as learning in high dimensional settings [34,35,65], learning without strong convexity assumptions [66], or relaxing the assumptions on data and objective functions [63, 70, 79].”
Bolt-on Differential Privacy for Scalable Stochastic Gradient Descent-based Analytics (2016, background): “Di ﬀ erentially private ML has been extensively studied by researchers from the database, ML, and theoretical computer science communities [10, 13, 15, 25, 27, 36, 37].”
Private Empirical Risk Minimization: Efficient Algorithms and Tight Error Bounds (2014, background): “Differentially private convex learning in different models has also been studied: for example, [21, 14, 38] study online optimization, [22] study an interactive model tailored to highdimensional kernel learning.”
Distributed Learning without Distress: Privacy-Preserving Empirical Risk Minimization (2018, background): “Many works have shown that differential privacy can be used to enable privacy-preserving machine learning in the centralized setting where a single organization owns all the data [10, 11, 29, 30, 51, 58].”