paper-with-me

홈 › Papers

Natural Gradient Methods: Perspectives, Efficient-Scalable Approximations, and Analysis

2023-03-06 · Rajesh Shrestha

Natural Gradient Descent, a second-degree optimization method motivated by the information geometry, makes use of the Fisher Information Matrix instead of the Hessian which is typically used. However, in many cases, the Fisher Information Matrix is equivalent to the Generalized Gauss-Newton Method, that both approximate the Hessian. It is an appealing method to be used as an alternative to stochastic gradient descent, potentially leading to faster convergence. However, being a second-order method makes it infeasible to be used directly in problems with a huge number of parameters and data. This is evident from the community of deep learning sticking with the stochastic gradient descent method since the beginning. In this paper, we look at the different perspectives on the natural gradient method, study the current developments on its efficient-scalable empirical approximations, and finally examine their performance with extensive experiments.

📄 PDF Abstract BibTeX arXiv:2303.05473

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Fast and Simple Natural-Gradient Variational Inference with Mixture of Exponential-family Approximations

2019-06-07 · Wu Lin, Mohammad Emtiyaz Khan, Mark Schmidt

Natural-gradient methods enable fast and simple algorithms for variational inference, but due to computational difficulties, their use is mostly limited to \emph{minimal} exponential-family (EF) approximations. In this p…

Bayesian InferenceVariational Inference

Understanding the Effects of Second-Order Approximations in Natural Policy Gradient Reinforcement Learning

2022-01-22 · Brennan Gebotys, Alexander Wong, David A. Clausi

Natural policy gradient methods are popular reinforcement learning methods that improve the stability of policy gradient methods by utilizing second-order approximations to precondition the gradient with the inverse of t…

Policy Gradient Methodsreinforcement-learningReinforcement LearningReinforcement Learning (RL)

Layerwise LQR for Geometry-Aware Optimization of Deep Networks

2026-05-05 · Simon Dufort-Labbé, Pierre-Luc Bacon, Razvan Pascanu, Simon Lacoste-Julien 외 arxiv

Geometry-aware optimizers such as Newton and natural gradient can improve conditioning in deep learning, but scalable variants such as K-FAC, Shampoo, and related preconditioners usually impose structural approximations …

A Coordinate-Free Construction of Scalable Natural Gradient

2018-08-30 · Kevin Luk, Roger Grosse

Most neural networks are trained using first-order optimization methods, which are sensitive to the parameterization of the model. Natural gradient descent is invariant to smooth reparameterizations because it is defined…

Scalable Gaussian-process regression and variable selection using Vecchia approximations

2022-02-25 · Jian Cao, Joseph Guinness, Marc G. Genton, Matthias Katzfuss

Gaussian process (GP) regression is a flexible, nonparametric approach to regression that naturally quantifies uncertainty. In many applications, the number of responses and covariates are both large, and a goal is to se…

regressionVariable Selection