English
Related papers

Related papers: Policy Gradient with Second Order Momentum

200 papers

In this paper, we study stochastic non-convex optimization with non-convex random functions. Recent studies on non-convex optimization revolve around establishing second-order convergence, i.e., converging to a nearly second-order optimal…

Optimization and Control · Mathematics 2017-11-02 Mingrui Liu , Tianbao Yang

Natural gradient descent is an optimization method traditionally motivated from the perspective of information geometry, and works well for many applications as an alternative to stochastic gradient descent. In this paper we critically…

Machine Learning · Computer Science 2020-09-22 James Martens

Momentum methods, such as Polyak's Heavy Ball, are the standard for training deep networks but suffer from curvature-induced bias in stochastic settings, limiting convergence to suboptimal $\mathcal{O}(\epsilon^{-4})$ rates. Existing…

Optimization and Control · Mathematics 2026-05-18 El Mahdi Chayti

Deep neural networks are usually trained with stochastic gradient descent (SGD), which minimizes objective function using very rough approximations of gradient, only averaging to the real gradient. Standard approaches like momentum or ADAM…

Machine Learning · Computer Science 2023-03-14 Jarek Duda

The success of deep learning can be attributed to various factors such as increase in computational power, large datasets, deep convolutional neural networks, optimizers etc. Particularly, the choice of optimizer affects the generalization,…

Machine Learning · Computer Science 2021-09-10 Anirudh Maiya , Inumella Sricharan , Anshuman Pandey , Srinivas K. S

Standard gradient descent methods are susceptible to a range of issues that can impede training, such as high correlations and different scaling in parameter space.These difficulties can be addressed by second-order approaches that apply a…

Machine Learning · Computer Science 2020-04-29 Ted Moskovitz , Rui Wang , Janice Lan , Sanyam Kapoor , Thomas Miconi , Jason Yosinski , Aditya Rawal

We present a novel approach to accelerate stochastic gradient descent (SGD) by utilizing curvature information obtained from Hessian-vector products or finite differences of parameters and gradients, similar to the BFGS algorithm. Our…

Machine Learning · Computer Science 2024-02-08 Omead Pooladzandi , Xi-Lin Li

Pruning remains an effective strategy for reducing both the costs and environmental impact associated with deploying large neural networks (NNs) while maintaining performance. Classical methods, such as OBD (LeCun et al., 1989) and OBS…

Machine Learning · Computer Science 2026-01-21 Ivo Gollini Navarrete , Nicolás Mauricio Cuadrado Ávila , Martin Takáč , Samuel Horváth

Reinforcement Learning, particularly through policy gradient methods, has played a central role in enabling reasoning capabilities of Large Language Models. However, the optimization stability of policy gradients in this setting remains…

Machine Learning · Computer Science 2026-03-03 Luckeciano C. Melo , Alessandro Abate , Yarin Gal

We propose a fast second-order method that can be used as a drop-in replacement for current deep learning solvers. Compared to stochastic gradient descent (SGD), it only requires two additional forward-mode automatic differentiation…

Machine Learning · Computer Science 2018-05-22 João F. Henriques , Sebastien Ehrhardt , Samuel Albanie , Andrea Vedaldi

We show that on-policy policy gradient (PG) and its variance reduction variants can be derived by taking finite difference of function evaluations supplied by estimators from the importance sampling (IS) family for off-policy evaluation…

Machine Learning · Computer Science 2020-06-25 Jiawei Huang , Nan Jiang

In the paper, we design a novel Bregman gradient policy optimization framework for reinforcement learning based on Bregman divergences and momentum techniques. Specifically, we propose a Bregman gradient policy optimization (BGPO) algorithm…

Machine Learning · Computer Science 2022-03-17 Feihu Huang , Shangqian Gao , Heng Huang

Second-order information -- such as curvature or data covariance -- is critical for optimisation, diagnostics, and robustness. However, in many modern settings, only the gradients are observable. We show that the gradients alone can reveal…

Machine Learning · Computer Science 2026-04-08 Arash Jamshidi , Katsiaryna Haitsiukevich , Kai Puolamäki

We introduce a novel algorithm for gradient-based optimization of stochastic objective functions. The method may be seen as a variant of SGD with momentum equipped with an adaptive learning rate automatically adjusted by an 'energy'…

Optimization and Control · Mathematics 2022-03-24 Hailiang Liu , Xuping Tian

This paper proposes a family of online second order methods for possibly non-convex stochastic optimizations based on the theory of preconditioned stochastic gradient descent (PSGD), which can be regarded as an enhance stochastic Newton…

Machine Learning · Statistics 2018-05-01 Xi-Lin Li

Second-order information is valuable for many applications but challenging to compute. Several works focus on computing or approximating Hessian diagonals, but even this simplification introduces significant additional costs compared to…

Machine Learning · Computer Science 2024-07-08 Mohamed Elsayed , Homayoon Farrahi , Felix Dangel , A. Rupam Mahmood

Natural gradients have long been studied in deep reinforcement learning due to their fast convergence properties and covariant weight updates. However, computing natural gradients requires inversion of the Fisher Information Matrix (FIM) at…

Machine Learning · Computer Science 2026-02-12 Yingxiao Huo , Satya Prakash Dash , Radu Stoican , Samuel Kaski , Mingfei Sun

We design an algorithm which finds an $\epsilon$-approximate stationary point (with $\|\nabla F(x)\|\le \epsilon$) using $O(\epsilon^{-3})$ stochastic gradient and Hessian-vector products, matching guarantees that were previously available…

Machine Learning · Computer Science 2020-06-25 Yossi Arjevani , Yair Carmon , John C. Duchi , Dylan J. Foster , Ayush Sekhari , Karthik Sridharan

First order methods, which solely rely on gradient information, are commonly used in diverse machine learning (ML) and data analysis (DA) applications. This is attributed to the simplicity of their implementations, as well as low…

Machine Learning · Computer Science 2018-03-06 Sudhir B. Kylasa , Farbod Roosta-Khorasani , Michael W. Mahoney , Ananth Grama

Momentum-based gradients are essential for optimizing advanced machine learning models, as they not only accelerate convergence but also advance optimizers to escape stationary points. While most state-of-the-art momentum techniques utilize…

Machine Learning · Computer Science 2025-05-20 Wei Zhang , Arif Hassan Zidan , Afrar Jahin , Yu Bao , Tianming Liu