English
Related papers

Related papers: Practical Quasi-Newton Methods for Training Deep N…

200 papers

We propose an L-BFGS optimization algorithm on Riemannian manifolds using minibatched stochastic variance reduction techniques for fast convergence with constant step sizes, without resorting to linesearch methods designed to satisfy Wolfe…

Optimization and Control · Mathematics 2017-05-23 Anirban Roychowdhury

We introduce a quasi-Newton method with block updates called Block BFGS. We show that this method, performed with inexact Armijo-Wolfe line searches, converges globally and superlinearly under the same convexity assumptions as BFGS. We also…

Optimization and Control · Mathematics 2017-12-04 Wenbo Gao , Donald Goldfarb

In this paper, we study and prove the non-asymptotic superlinear convergence rate of the Broyden class of quasi-Newton algorithms which includes the Davidon--Fletcher--Powell (DFP) method and the Broyden--Fletcher--Goldfarb--Shanno (BFGS)…

Optimization and Control · Mathematics 2021-12-02 Qiujiang Jin , Aryan Mokhtari

We extend the well-known BFGS quasi-Newton method and its memory-limited variant LBFGS to the optimization of nonsmooth convex objectives. This is done in a rigorous fashion by generalizing three components of BFGS to subdifferentials: the…

Machine Learning · Statistics 2010-11-30 Jin Yu , S. V. N. Vishwanathan , Simon Guenter , Nicol N. Schraudolph

Despite the predominant use of first-order methods for training deep learning models, second-order methods, and in particular, natural gradient methods, remain of interest because of their potential for accelerating training through the use…

Machine Learning · Computer Science 2021-12-23 Yi Ren , Donald Goldfarb

Non-asymptotic convergence analysis of quasi-Newton methods has gained attention with a landmark result establishing an explicit local superlinear rate of O$((1/\sqrt{t})^t)$. The methods that obtain this rate, however, exhibit a well-known…

Optimization and Control · Mathematics 2023-10-19 Zhan Gao , Aryan Mokhtari , Alec Koppel

First-order methods such as stochastic gradient descent (SGD) are currently the standard algorithm for training deep neural networks. Second-order methods, despite their better convergence rate, are rarely used in practice due to the…

Machine Learning · Computer Science 2019-09-26 Tianle Cai , Ruiqi Gao , Jikai Hou , Siyu Chen , Dong Wang , Di He , Zhihua Zhang , Liwei Wang

This paper considers consensus optimization problems where each node of a network has access to a different summand of an aggregate cost function. Nodes try to minimize the aggregate cost function, while they exchange information only with…

Optimization and Control · Mathematics 2016-03-24 Mark Eisen , Aryan Mokhtari , Alejandro Ribeiro

We consider the use of a curvature-adaptive step size in gradient-based iterative methods, including quasi-Newton methods, for minimizing self-concordant functions, extending an approach first proposed for Newton's method by Nesterov. This…

Optimization and Control · Mathematics 2018-08-13 Wenbo Gao , Donald Goldfarb

Using quasi-Newton methods in stochastic optimization is not a trivial task given the difficulty of extracting curvature information from the noisy gradients. Moreover, pre-conditioning noisy gradient observations tend to amplify the noise.…

Optimization and Control · Mathematics 2024-04-02 Andre Carlon , Luis Espath , Raul Tempone

In this paper, we propose a second order optimization method to learn models where both the dimensionality of the parameter space and the number of training samples is high. In our method, we construct on each iteration a Krylov subspace…

Machine Learning · Statistics 2011-11-21 Oriol Vinyals , Daniel Povey

The gradient descent (GD) method has been used widely to solve parameter estimation in generalized linear models (GLMs), a generalization of linear models when the link function can be non-linear. In GLMs with a polynomial link function, it…

Optimization and Control · Mathematics 2024-03-15 Qiujiang Jin , Tongzheng Ren , Nhat Ho , Aryan Mokhtari

We present two sampled quasi-Newton methods (sampled LBFGS and sampled LSR1) for solving empirical risk minimization problems that arise in machine learning. Contrary to the classical variants of these methods that sequentially build…

Optimization and Control · Mathematics 2021-07-29 Albert S. Berahas , Majid Jahani , Peter Richtárik , Martin Takáč

In the context of deep learning, many optimization methods use gradient covariance information in order to accelerate the convergence of Stochastic Gradient Descent. In particular, starting with Adagrad, a seemingly endless line of research…

Machine Learning · Computer Science 2020-12-08 Nikolaos Tselepidis , Jonas Kohler , Antonio Orvieto

In this paper, we implement the Stochastic Damped LBFGS (SdLBFGS) for stochastic non-convex optimization. We make two important modifications to the original SdLBFGS algorithm. First, by initializing the Hessian at each step using an…

Machine Learning · Computer Science 2018-05-08 Yingkai Li , Huidong Liu

The natural gradient is central in neural quantum states optimizations but it is limited by the cost of computing and inverting the quantum geometric tensor, the quantum analogue of the Fisher information matrix. We introduce a…

Quantum Physics · Physics 2025-11-05 Ahmedeo Shokry , Alessandro Santini , Filippo Vicentini

In this paper, a novel stochastic extra-step quasi-Newton method is developed to solve a class of nonsmooth nonconvex composite optimization problems. We assume that the gradient of the smooth part of the objective function can only be…

Optimization and Control · Mathematics 2019-10-22 Minghan Yang , Andre Milzarek , Zaiwen Wen , Tong Zhang

In this paper, we present a novel approach to approximate the gain function of the feedback particle filter (FPF). The exact gain function is the solution of a Poisson equation involving a probability-weighted Laplacian. The numerical…

Machine Learning · Computer Science 2022-06-07 S. Yagiz Olmez , Amirhossein Taghvaei , Prashant G. Mehta

We present Exact Gauss-Newton (EGN), a stochastic second-order optimization algorithm that combines the generalized Gauss-Newton (GN) Hessian approximation with low-rank linear algebra to compute the descent direction. Leveraging the…

Machine Learning · Computer Science 2025-10-16 Mikalai Korbit , Adeyemi D. Adeoye , Alberto Bemporad , Mario Zanon

Deep Gaussian processes (DGPs) are multi-layer hierarchical generalisations of Gaussian processes (GPs) and are formally equivalent to neural networks with multiple, infinitely wide hidden layers. DGPs are nonparametric probabilistic models…