Heavy Ball method speeds up finding global optima in non-convex problems.
problem Finding global optima in non-convex optimization problems.
method Heavy Ball momentum in non-convex optimization.
result Heavy Ball helps iterates enter a benign region faster, containing a global optimal point.
Heavy-ball algorithms can always avoid saddle points with random initialization.
problem Optimizing nonconvex functions with saddle points.
method Developed a new mapping to interpret heavy-ball algorithms as iterations, proving they can escape saddle points.
result Heavy-ball algorithms can escape saddle points with random initialization.
This paper deals with a natural stochastic optimization procedure derived from the so-called Heavy-ball method differential equation, which was introduced by Polyak in the 1960s with his seminal contribution [Pol64]. The Heavy-ball method is a second-order dynamics that was investigated to minimize convex functions f .…
In this work we establish the first linear convergence result for the stochastic heavy ball method. The method performs SGD steps with a fixed stepsize, amended by a heavy ball momentum term. In the analysis, we focus on minimizing the expected loss and not on finite-sum minimization, which is typically a much harder p…
New method shows stochastic momentum can converge quickly on optimization problems.
problem Improving convergence of stochastic optimization methods.
method Stochastic heavy ball momentum with minibatching.
result Stochastic heavy ball momentum retains fast linear rate on quadratic problems.
In this paper, we revisit the convergence of the Heavy-ball method, and present improved convergence complexity results in the convex setting. We provide the first non-ergodic O(1/k) rate result of the Heavy-ball algorithm with constant step size for coercive objective functions. For objective functions satisfying a re…
Arguably, the two most popular accelerated or momentum-based optimization methods in machine learning are Nesterov's accelerated gradient and Polyaks's heavy ball, both corresponding to different discretizations of a particular second order differential equation with friction. Such connections with continuous-time dyna…
Gradient-based optimization algorithms can be studied from the perspective of limiting ordinary differential equations (ODEs). Motivated by the fact that existing ODEs do not distinguish between two fundamentally different algorithms---Nesterov's accelerated gradient method for strongly convex functions (NAG-SC) and Po…
Paper analyzes SHB method for neural networks, proving stability, connectivity, and global convergence.
problem Theoretical understanding of SHB method for neural networks.
method Mean-field analysis of SHB dynamics related to a partial differential equation.
result SHB method converges to global optimum and exhibits stability and connectivity.
A new Bayesian filtering method speeds up stochastic Newton optimization.
problem Minimizing log-convex functions using stochastic methods.
method Contextualizes the problem as Bayesian inference, applying Bayesian filtering to update estimates.
result Establishes conditions for diminishing effect of older observations, akin to momentum.
Two new differentially private optimization algorithms derived from accelerated methods.
problem Improving privacy in optimization algorithms while maintaining convergence rates.
method Polyak's heavy ball method and Nesterov's accelerated gradient method with differential privacy.
result The proposed algorithms outperform existing differentially private optimization methods.
The paper analyzes convergence rates for SGD and SHB methods.
problem Analyzing convergence rates for stochastic gradient descent and heavy ball methods.
method Stochastic gradient descent and stochastic heavy ball method for general stochastic approximation problems.
result The last iterate of SHB converges almost surely to a minimizer and has faster convergence rates than SGD.
Proposes momentum methods for Lie groups, improving on classical algorithms.
problem Optimization on nonlinear spaces, especially Lie groups.
method Generalizes Nesterov's Accelerated Gradient method to Lie groups.
result Demonstrates faster convergence for NAG-like methods on Lie groups.
This paper analyzes two Lie group momentum optimization algorithms and their convergence rates.
problem Optimizing functions on Lie groups using momentum-based dynamics.
method Investigates Lie Heavy-Ball and Lie NAG-SC algorithms, quantifying their convergence rates under smoothness and convexity assumptions.
result Lie NAG-SC accelerates optimization over the momentumless case, while Lie Heavy-Ball does not.
Study accelerates optimization methods in non-convex problems, but doesn't improve the algorithm's performance.
problem Understanding the behavior of momentum-based acceleration methods in non-convex, high-dimensional landscapes.
method Used dynamical mean field theory to describe the average dynamics of heavy-ball momentum and Nesterov acceleration in a non-convex model.
result Accelerated dynamics but did not improve the algorithm's performance with respect to gradient descent.
Paper proves SHB convergence with biased gradients and approximate step sizes.
problem Establishing convergence of SHB with biased gradients and approximate step sizes.
method Generalizes SHB convergence conditions for biased gradients, approximate step sizes, and block updating.
result Proves convergence of SHB with new conditions for biased gradients and approximate step sizes.
Improved DANE algorithm for faster convergence in distributed machine learning.
problem Challenges in convergence of DANE algorithm for general convex functions.
method Introducing variants of DANE with backtracking line search and heavy-ball method.
result Proved global and local convergence rates for quadratic and non-quadratic strongly convex functions.
Recently, {\it stochastic momentum} methods have been widely adopted in training deep neural networks. However, their convergence analysis is still underexplored at the moment, in particular for non-convex optimization. This paper fills the gap between practice and theory by developing a basic convergence analysis of t…
A novel decentralized deep learning algorithm using gradient-based optimization.
problem Decentralized deep learning in networked systems without a central server.
method Heavy-ball acceleration method and consensus protocol for model and gradient-momentum sharing.
result The proposed algorithm outperforms competing methods in various communication topologies.
In this paper we study several classes of stochastic optimization algorithms enriched with heavy ball momentum. Among the methods studied are: stochastic gradient descent, stochastic Newton, stochastic proximal point and stochastic dual subspace ascent. This is the first time momentum variants of several of these metho…
Optimizes web page freshness with limited crawling frequencies.
problem Maximize local cache freshness given crawling frequency constraints.
method Three novel online estimation schemes for page change rates.
result Convergent algorithms for estimating page change rates.
Improved SHB method for faster convergence on strongly-convex quadratics.
problem Understanding and improving the theoretical and practical advantages of SHB.
method Noise-adaptive multi-stage algorithm for SHB with accelerated convergence.
result SHB can achieve accelerated convergence with larger mini-batch sizes.
Optimization algorithms help overparameterized neural networks achieve high performance.
problem Understanding the convergence of optimization algorithms on overparameterized neural networks.
method Analyzing a broad class of optimization algorithms using dynamical systems and finite over-parameterized neural networks with ReLU activation.
result The Heavy Ball method converges to global minimum at a linear rate, while NAG converges sublinearly.
We study first-order optimization methods obtained by discretizing ordinary differential equations (ODEs) corresponding to Nesterov's accelerated gradient methods (NAGs) and Polyak's heavy-ball method. We consider three discretization schemes: an explicit Euler scheme, an implicit Euler scheme, and a symplectic scheme.…
Study accelerates gradient methods in machine learning, revealing risk and stability connections.
problem Understanding the statistical risk of accelerated gradient methods in machine learning.
method Continuous-time analysis of Nesterov's accelerated gradient method and Polyak's heavy ball method for least squares regression.
result Connections between early stopping, stability, and curvature of loss function are revealed.
Generalizes momentum methods using Hamiltonian dynamics.
problem Optimization in constrained Euclidean and non-Euclidean spaces.
method Hamiltonian perspective to generalize momentum methods.
result Generic and unifying nonasymptotic analysis of convergence.
Fine-grained analysis of gradient descent with momentum provides modified loss equations.
problem Understanding the dynamics of gradient descent with momentum.
method Fine-grained analysis and derivation of modified loss equations.
result Global approximation bounds and continuous modified equations for HB.
Adaptive gradient methods such as Adam have been shown to be very effective for training deep neural networks (DNNs) by tracking the second moment of gradients to compute the individual learning rates. Differently from existing methods, we make use of the most recent first moment of gradients to compute the individual …
Acceleration in Hilbert spaces reduces computations but not accuracy.
problem Improving learning accuracy with fewer computations.
method Analysis of Nesterov acceleration and heavy-ball methods in Hilbert spaces.
result Acceleration can reduce computations but not improve accuracy with respect to gradient descent.
We show that accelerated gradient descent, averaged gradient descent and the heavy-ball method for non-strongly-convex problems may be reformulated as constant parameter second-order difference equation algorithms, where stability of the system is equivalent to convergence at rate O(1/n 2), where n is the number of ite…
Derives new optimization methods using variational integrators.
problem Optimization methods in machine learning.
method Variational integrators and principles of Hamilton and Lagrange-d'Alembert.
result Derives two families of optimization methods, including Nesterov's accelerated gradient method.
The paper analyzes how momentum affects convergence in stochastic gradient methods.
problem Lack of clear understanding of momentum's impact on convergence and performance.
method Unified analysis of several popular algorithms using the QHM formulation.
result Provides practical guidelines for setting learning rate and momentum parameters.
Extends random feature analysis to spectral methods and improves learning rates.
problem Improving generalization properties of spectral methods in large-scale learning.
method Extends random feature analysis to a broad class of spectral regularization techniques, including gradient descent and Nesterov method.
result Obtains optimal learning rates for regularity classes, including those not in the RKHS.
There is widespread sentiment that it is not possible to effectively utilize fast gradient methods (e.g. Nesterov's acceleration, conjugate gradient, heavy ball) for the purposes of stochastic optimization due to their instability and error accumulation, a notion made precise in d'Aspremont 2008 and Devolder, Glineur, …
Stochastic momentum methods have been widely adopted in training deep neural networks. However, their theoretical analysis of convergence of the training objective and the generalization error for prediction is still under-explored. This paper aims to bridge the gap between practice and theory by analyzing the stochast…
Two major momentum-based techniques that have achieved tremendous success in optimization are Polyak's heavy ball method and Nesterov's accelerated gradient. A crucial step in all momentum-based methods is the choice of the momentum parameter m which is always suggested to be set to less than 1. Although the choice…
New method improves optimization and DP in FL.
problem Combining strong DP and optimization in FL.
method Combining clipping, momentum, and error feedback.
result Optimal convergence rate and near optimal DP guarantees.
Rod flow models Adam's behavior at the edge of stability.
problem Modeling adaptive gradient methods like Adam at the edge of stability.
method Extended rod flow to Adam, considering parameters, first moment, and second moment as variables.
result Rod flow accurately tracks Adam's behavior through the edge-of-stability regime.
A method for estimating the median of gradients in stochastic optimization.
problem Robust gradient estimation in stochastic optimization for various applications.
method Stochastic Proximal Point Method for median gradient estimation.
result The proposed method can converge even under heavy-tailed, state-dependent noise.
Analysis of momentum methods on quadratic models, showing SGD's superiority.
problem Analysis of stochastic gradient algorithms with momentum on quadratic models.
method Inspired by random matrix theory, exact characterization of loss values.
result Stochastic heavy-ball momentum does not improve over SGD in the strongly convex setting.
A new method corrects staleness in online optimization by transporting past gradients.
problem Reducing staleness and variance in online optimization methods.
method Implicit gradient transport (IGT) to correct past gradients at the current iterate.
result IGT reduces variance and bias in updates over time and achieves state-of-the-art results.
Improved convergence for nonconvex optimization with dependent data.
problem Constrained smooth nonconvex optimization with dependent data.
method Stochastic projected gradient methods under a general dependent data sampling scheme.
result Achieved worst-case rate of convergence ildeO(t−1/4) and complexity ildeO(ε−4). Paper studies stochastic optimization methods with momentum, proving convergence and avoiding traps.
problem Optimizing non-convex functions with momentum.
method Unified analysis of stochastic gradient descent variants, including S-NAG and Adam.
result Convergence to critical points and avoidance of undesired critical points like local maxima or saddle points.
First order optimization algorithms play a major role in large scale machine learning. A new class of methods, called adaptive algorithms, were recently introduced to adjust iteratively the learning rate for each coordinate. Despite great practical success in deep learning, their behavior and performance on more genera…
Optimizes convex functions in finite vs infinite dimensions, revealing slow convergence rates.
problem Analyzing gradient flows in finite and infinite-dimensional Hilbert spaces.
method Proves convergence rates and optimality conditions for gradient flows and related methods.
result Gradient flow convergence rates in finite dimensions are slower than in infinite dimensions, with optimal rates achievable in Hilbert spaces.
Standard optimizers perform as well as LARS and LAMB at large batch sizes.
problem Comparing optimizers for neural network training at large batch sizes.
method Used standard optimizers like Nesterov momentum and Adam to match or exceed LARS and LAMB results.
result Standard optimizers can match or exceed LARS and LAMB at large batch sizes.
While momentum-based accelerated variants of stochastic gradient descent (SGD) are widely used when training machine learning models, there is little theoretical understanding on the generalization error of such methods. In this work, we first show that there exists a convex loss function for which the stability gap fo…
Momentum based stochastic gradient methods such as heavy ball (HB) and Nesterov's accelerated gradient descent (NAG) method are widely used in practice for training deep networks and other supervised learning models, as they often provide significant improvements over stochastic gradient descent (SGD). Rigorously speak…