Large-Step-Size Gradient Descent
Latest papers
10 papers2 papers in the last four weeks, against 1 the four weeks before. 0.0% of all new papers.
Neural networks are often trained with gradient descent (GD) using step sizes that exceed classical stability thresholds, a regime where standard convergence theory no longer applies. Despite exhibiting non-monotonic loss and oscillatory dynamics, this regime frequently leads to faster convergence and better generalization. Existing theoretical perspectives offer complementary but incomplete insights: one identifies a self-stabilization mechanism that drives GD toward flat regions but does not yield convergence rates; another derives rigorous rates for specific losses but does not explain the underlying stabilization mechanism; and a third unifies these views for overparameterized least squares, but relies on the existence of a minimizer manifold, which is absent in classification settings. In this work, we extend this unified geometric perspective to exponential-tail losses, including logistic and cross-entropy losses in separable linear classification, where the loss infimum is not attained at any finite parameter. We identify a curvature-defined reference subspace that replaces the role of the minimizer manifold, yielding a coupled dynamical system in which orthogonal oscillations are damped by parallel progress along directions of decreasing sharpness. Under regularity conditions, which we verify for logistic and multiclass cross-entropy losses on symmetric data models, we derive explicit bounds for the iterates, not just the loss, covering both the unstable and stable regimes, and revealing a three-phase damping structure in the latter. Our analysis further uncovers a transient implicit bias induced by large step sizes: although GD converges asymptotically to the max-margin direction, large steps can leave persistent orthogonal residuals and, in certain geometries, amplify them, affecting finite-time alignment with the max-margin direction.
Adversarial Training with Large Step Sizes: Implicit Bias, Optimization, and Sharpness
Adversarial Training (AT) is an important method for improving model robustness. We study the dynamics of deterministic Gradient Descent (GD) in AT. Most existing works in this direction focus on the small or vanishing step sizes regime, where GD resembles a continuous-time flow. However, such step sizes can make training inefficient and are rarely used in practice. In this paper, we study the AT dynamics of GD with large constant step sizes. We first study GD in adversarial linear logistic regression, focusing on two aspects: where the model parameters converge (implicit bias) and how quickly the losses decrease (optimization). We then turn to neural networks training and study how the sharpness of losses evolve during AT. Our results advance the understanding of the behaviors of GD with large step sizes in AT and highlight three key differences from its behavior in Standard Training (ST): (i) the implicit bias gap between AT and ST, previously established only for small step sizes, persists for arbitrarily large constant step sizes, (ii) different from ST, there is no uniform rate of convergence for the optimization error in AT, and (iii) sharpness can exhibit two-stage dynamics that differ from the Edge-of-Stability behavior observed in ST.
Gradient descent reaches arbitrary accuracy in a fixed number of large steps on multiclass logistic regression
We show that gradient descent (GD) with step size can achieve optimality gap on the multiclass logistic regression loss in a number of iterations that is independent of . To do so, the step size must be chosen large enough as a function of (and properties of the linearly separable dataset), while is a constant that only depends on the data. The relevant properties of the data include the number of points and their margin (for ), and one additional quantity (for ) that turns out to be necessary in dimension . The proofs build on the known fact that, as , suitably rescaled GD becomes the perceptron algorithm, and the latter finds a perfect classifier in iterations. A similar claim was shown for two classes in dimension using other proof techniques. This result adds to previous works that showed GD on the logistic regression loss can be accelerated by taking large step sizes. Those provoke a wild, oscillatory initial phase, followed by a stable phase. We provide results of this (qualitatively different) nature, too.
Improved Convergence of Large Stepsize Gradient Descent for Logistic Regression
We study gradient descent (GD) with a large constant stepsize for logistic regression on linearly separable data. Existing analysis shows an accelerated rate of to reach loss with an aggressive stepsize, although the loss may initially oscillate. Tighter control of the oscillatory dynamics has been available only for two-dimensional data. We prove a substantially faster rate in arbitrary dimension: GD with a large stepsize reaches loss within steps, where depends only on the margin and the rank of the data. Our proof improves the bound on the transition time of GD from the oscillatory to the stable phase, after which the loss decreases monotonically. We split the oscillatory phase into recursively nested intervals. The margin and the rank bound the nesting depth, and a counting argument bounds the number of intervals at each depth, together yielding the polylogarithmic step complexity.
Dynamics of Gradient Descent with Large Step Size Near a Manifold of Flat Minima
An important quantity in the theory of gradient descent (GD) is the sharpness, defined as the largest eigenvalue of the objective Hessian. Classical analyses typically require the step size to be uniformly smaller than twice the reciprocal of the sharpness, but this condition is frequently violated in the training of deep neural networks. Recent work bridges this gap in the setting of overparametrised least-squares with a single scalar output, providing a normal form for large-step GD in a neighbourhood of an isolated flat minimum and establishing three corresponding convergence results. In this paper, we extend this theory in two directions: (1) to overparametrised least-squares with vector-valued outputs (including regression with arbitrarily many observations), and (2) to a neighbourhood of a manifold of flat minima (which we show is essential for applications such as matrix factorisation). We generalise both the normal form and all three convergence theorems of \cite{macdonaldeos} to this broader setting, overcoming several technical challenges, including the solution of a singular partial differential equation via a novel method that may be of independent interest. We further show that our framework applies to deep matrix factorisation under mild assumptions, yielding several new structural results. In particular, we prove that the set of flat minima forms a fibre bundle over a product of spheres, and that the sharpness is Morse-Bott along this manifold.
Flatland: The Adventures of Gradient Descent with Large Step Sizes
The training of neural networks often entails objective functions that are not globally -smooth. For these functions, it is both theoretically and practically difficult to reply to the question: what is the largest possible step size that ensures the convergence of gradient descent (GD)? We address this longstanding open question in deep learning by providing a unifying definition of "large" step sizes that requires only local Lipschitz (or even Hölder) continuity of the gradient. We design first-order adaptive methods that provably yield large step sizes and show that they operate at the edge of stability (EoS) right from the start of the training. In particular, the loss decreases nonmonotonically and the product between the step size and sharpness, i.e., the largest eigenvalue of the Hessian, stays above the EoS threshold of 2 throughout training. Using our method, we are also able to minimize the sharpness all the way down to its global minimum. Contrary to expectation, we find that encountering globally-flat regions too early in the training may both slow down convergence and jeopardize the generalization ability of the network. Exploiting a self-stabilization argument, we allow GD to enter slightly sharper valleys and turn unsuccessful training runs into very successful ones.
Locally invariant manifolds and foliations
Invariant manifolds and foliations associated to invariant pseudo-hyperbolic reference geometries have been well-studied in the context of dynamics implemented by a diffeomorphism. In this paper, we provide sufficient conditions for the existence of maximally smooth, locally invariant manifolds and foliations containing a given locally invariant, pseudo-hyperbolic reference geometry, in the context of dynamics implemented by a smooth map which need not be invertible. The theory is motivated by applications to the developing theory of large-step optimisation.
Gradient Descent on Logistic Regression: Do Large Step-Sizes Work with Data on the Sphere?
Gradient descent on logistic regression has many fascinating properties not fully captured by classical analyses. For linearly separable data, the iterates converge in direction to the maximum-margin separator regardless of how large the step size is. For non-separable data, however, cycles may occur even for step sizes below the local stability threshold of , where is the largest eigenvalue of the Hessian at the solution. We ask whether requiring all examples to have equal norm is a sufficient condition for global convergence below this threshold. We prove that this is indeed true in one dimension. In higher dimensions, even equal-norm datasets can produce cycles, although global convergence holds for the population loss when the examples are uniformly distributed on the sphere. Finally, in one dimension, we relax the equal-norm assumption to an annulus condition that ensures global convergence below the stability threshold, suggesting a path towards analogous sufficient conditions in higher dimensions.
Bias-Optimal Bounds for Stochastic Gradient Descent
The non-asymptotic analysis of Stochastic Gradient Descent (SGD) typically yields bounds that decompose into a bias term and a variance term. In this work, we focus on the bias component and study the extent to which SGD can match the optimal convergence behavior of deterministic gradient descent. Assuming only (strong) convexity and smoothness of the objective, we derive new bounds for the full range of constant step-sizes . This includes critical and large step-size regimes that were previously unexplored without additional variance assumptions, in both the convex and strongly convex cases. Those bounds are bias-optimal, in the sense that the bias term coincides with the worst-case rate of gradient descent, for all non-critical step-sizes. For critical step-sizes, we show that no bounds can be obtained with an optimal bias term, but that this optimal bias can be approximated arbitrarily closely. As an immediate application, we provide new complexity bounds for the Stochastic Proximal algorithm, without the need of regularity assumptions.
Gradient Descent on Logistic Regression with Non-Separable Data and Large Step Sizes
We study gradient descent (GD) dynamics on logistic regression problems with large, constant step sizes. For linearly-separable data, it is known that GD converges to the minimizer with arbitrarily large step sizes, a property which no longer holds when the problem is not separable. In fact, the behaviour can be much more complex -- a sequence of period-doubling bifurcations begins at the critical step size , where is the largest eigenvalue of the Hessian at the solution. Using a smaller-than-critical step size guarantees convergence if initialized nearby the solution: but does this suffice globally? In one dimension, we show that a step size less than suffices for global convergence. However, for all step sizes between and the critical step size , one can construct a dataset such that GD converges to a stable cycle. In higher dimensions, this is actually possible even for step sizes less than . Our results show that although local convergence is guaranteed for all step sizes less than the critical step size, global convergence is not, and GD may instead converge to a cycle depending on the initialization.