Differentially private Riemannian optimization
원문 초록
In this paper, we study the differentially private empirical risk minimization problem where the parameter is constrained to a Riemannian manifold. We introduce a framework of differentially private Riemannian optimization by adding noise to the Riemannian gradient on the tangent space. The noise follows a Gaussian distribution intrinsically defined with respect to the Riemannian metric. We adapt the Gaussian mechanism from the Euclidean space to the tangent space compatible to such generalized Gaussian distribution. We show that this strategy presents a simple analysis as compared to directly adding noise on the manifold. We further show privacy guarantees of the proposed differentially private Riemannian (stochastic) gradient descent using an extension of the moments accountant technique. Additionally, we prove utility guarantees under geodesic (strongly) convex, general nonconvex objectives as well as under the Riemannian Polyak-{\L}ojasiewicz condition. We show the efficacy of the proposed framework in several applications.
함께 살펴볼 논문
공통 태그를 근거로 연결했습니다. 실제 인용 관계나 논문의 우수성을 뜻하지 않습니다.
Efficient Optimization with Orthogonality Constraint: a Randomized Riemannian Submanifold Method
Riemannian Optimization on Relaxed Indicator Matrix Manifold
Causal Abstraction Learning based on the Semantic Embedding Principle
A Spectral Framework for Tracking Communities in Evolving Networks
Elliptical Wishart distributions: information geometry, maximum likelihood estimator, performance analysis and statistical learning
The Gaussian Discriminant Variational Autoencoder (GdVAE): A Self-Explainable Model with Counterfactual Explanations
토론
이 논문과 연결된 포럼 글이 아직 없습니다.