Searcharxiv⌕ Search

arXiv · 2611.20437

On the Convergence of Continuous-Time AdamW

Abstract

Continuous-time AdamW provides a framework for understanding whether an optimizer widely used to train large-scale neural networks settles to an equilibrium or sustains persistent motion. We study this long-time behavior while retaining bias correction and fixed positive weight decay. Under suitable assumptions, we establish that the dynamics are well defined for all time and remain bounded when the decay is sufficiently small. Our main result connects their limiting behavior to the geometry of the objective: trajectories approach a locally invariant manifold whose dimension equals the Hessian nullity along the associated critical component. When critical points are isolated and have at most one such direction, the trajectory converges to a single equilibrium. With two degenerate directions, however, bounded persistent motion can occur even for a smooth, strictly convex objective with a unique minimum. We also quantify how close the long-time trajectory remains to the critical set as weight decay decreases, without assuming convergence to a point. These results identify objective geometry as a key to understanding when continuous-time AdamW converges and when persistent motion is possible.

Explore related subjects

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Bangjie Wang, Ning Ning. 2026-11-16. On the Convergence of Continuous-Time AdamW. https://searcharxiv.com/abs/2611.20437

Citations retain the original authors and DOI and include this site’s record URL. If this discovery service contributed to your research, you can also cite Searcharxiv by Indiciae: BibTeX · RIS.

KEEP EXPLORING

Related papers

First-order methods for stochastic and finite-sum convex optimization with deterministic constraints

In this paper, we study a class of stochastic and finite-sum convex optimization problems with deterministic constraints. Existing methods typically aim to find an $ε$-$expected\ feasible\ stochastic\ optimal$ solution, in which both the expected constraint violation and the expected optimality gap are within a prescribed tolerance $ε$. However, in many practical applications, constraints must be nearly satisfied with certainty, rendering such solutions potentially unsuitable due to the risk of substantial violations. To address this issue, we propose stochastic first-order methods for finding an $ε$-$surely\ feasible\ stochastic\ optimal$ ($ε$-SFSO) solution, where the constraint violation is deterministically bounded by $ε$ and the expected optimality gap is at most $ε$. Our methods apply an accelerated stochastic gradient (ASG) scheme or a modified variance-reduced ASG scheme $only\ once$ to a sequence of quadratic penalty subproblems with appropriately chosen penalty parameters. We establish first-order oracle complexity bounds for the proposed methods in computing an $ε$-SFSO solution. Preliminary numerical results are presented to illustrate the performance of the proposed methods.

math.OC↗

Attention's forward pass and Frank-Wolfe

We study the hardmax limit of self-attention dynamics for token embeddings obtained in the zero-temperature ($β\to+\infty$) regime, and relate it to the finite-$β$ setting. In this limit, the update rule can be viewed as a Frank-Wolfe step for a quadratic objective over the convex hull of the current token embeddings. When the key-query matrix is negative semidefinite, the method linearly contracts all tokens to a single cluster at the origin. When it is positive semidefinite, extending the hardmax rule to the entire convex hull induces a Voronoi diagram: vertices are stationary, interior points remain in their initial cells, and each token moves along a straight line toward its cell's vertex, yielding (super-)exponential convergence. As a byproduct, we also establish well-posedness of the associated ODE limit in this regime. Returning to the finite-$β$ regime, we model self-attention dynamics as a Markov chain and prove dynamic metastability: with high probability, interior tokens reach near-vertex configurations in a constant number of steps and remain within a small neighborhood for times that grow exponentially in the inverse temperature $β$, before ultimately collapsing to the origin. Thus, the hardmax dynamics accurately approximate the finite-$β$ process over exponentially long time horizons.

math.OC↗

Finite de Finetti for convex bodies and Polynomial Optimization

We study polynomial optimization over convex bodies with separately affine objectives and local affine equality and inequality constraints. We construct a convergent hierarchy of conic relaxations with explicit finite-level error bounds of order $n^{-1/2}$ at level $n$, complemented by a constructive rounding scheme that produces feasible points. The relaxation bounds and rounded objective values bracket the optimum at the same rate. The analytic ingredient is a finite de Finetti representation theorem for arbitrary proper cones: bipartite marginals of symmetric extensions in maximal tensor products are $O(n^{-1/2})$-close, in base norm, to points of the minimal tensor product. Its proof combines a uniform, monogamy-type bound on a variational mutual information based on Frenkel's relative entropy with classical conditioning and informationally complete measurements. We apply the framework to the assemblage relaxation of two-player non-local games in general probabilistic theories. For polytopic state spaces, the hierarchy consists of linear programs, and experiments on the Clauser--Horne--Shimony--Holt (CHSH) game provide a numerical proof of concept for the outer approximations and feasible rounding procedure.

math.OC↗