Title | ||
---|---|---|
Revisiting The Role Of Euler Numerical Integration On Acceleration And Stability In Convex Optimization |
Abstract | ||
---|---|---|
Viewing optimization methods as numerical integrators for ordinary differential equations (ODEs) provides a thought-provoking modern framework for studying accelerated first-order optimizers. In this literature, acceleration is often supposed to be linked to the quality of the integrator (accuracy, energy preservation, symplecticity). In this work, we propose a novel ordinary differential equation that questions this connection: both the explicit and the semi-implicit (a.k.a symplectic) Euler discretizations on this ODE lead to an accelerated algorithm for convex programming. Although semi-implicit methods are well-known in numerical analysis to enjoy many desirable features for the integration of physical systems, our findings show that these properties do not necessarily relate to acceleration. |
Year | Venue | DocType |
---|---|---|
2021 | 24TH INTERNATIONAL CONFERENCE ON ARTIFICIAL INTELLIGENCE AND STATISTICS (AISTATS) | Conference |
Volume | ISSN | Citations |
130 | 2640-3498 | 0 |
PageRank | References | Authors |
0.34 | 0 | 5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Peiyuan Zhang | 1 | 0 | 0.34 |
Orvieto, Antonio | 2 | 0 | 3.04 |
Hadi Daneshmand | 3 | 10 | 1.16 |
Thomas Hofmann | 4 | 10064 | 1001.83 |
Roy Smith | 5 | 0 | 0.34 |