Matches in SemOpenAlex for { <https://semopenalex.org/work/W4315589177> ?p ?o ?g. }
Showing items 1 to 61 of
61
with 100 items per page.
- W4315589177 abstract "Although Nesterov's accelerated gradient (NAG) methods have been studied from various perspectives, it remains unclear why the most popular forms of NAG must handle convex and strongly convex objective functions separately. Motivated by this inconsistency, we propose an NAG method that unifies the existing ones for the convex and strongly convex cases. We first design a Lagrangian function that continuously extends the first Bregman Lagrangian to the strongly convex setting. As a specific case of the Euler--Lagrange equation for this Lagrangian, we derive an ordinary differential equation (ODE) model, which we call the unified NAG ODE, that bridges the gap between the ODEs that model NAG for convex and strongly convex objective functions. We then design the unified NAG, a novel momentum method whereby the continuous-time limit corresponds to the unified ODE. The coefficients and the convergence rates of the unified NAG and unified ODE are continuous in the strong convexity parameter $mu$ on $[0, +infty)$. Unlike the existing popular algorithm and ODE for strongly convex objective functions, the unified NAG and the unified NAG ODE always have superior convergence guarantees compared to the known algorithms and ODEs for non-strongly convex objective functions. This property is beneficial in practical perspective when considering strongly convex objective functions with small $mu$. Furthermore, we extend our unified dynamics and algorithms to the higher-order setting. Last but not least, we propose the unified NAG-G ODE, a novel ODE model for minimizing the gradient norm of strongly convex objective functions. Our unified Lagrangian framework is crucial in the process of constructing this ODE. Fascinatingly, using our novel tool, called the differential kernel, we observe that the unified NAG ODE and the unified NAG-G ODE have an anti-transpose relationship." @default.
- W4315589177 created "2023-01-11" @default.
- W4315589177 creator A5063507594 @default.
- W4315589177 creator A5075218521 @default.
- W4315589177 date "2023-01-09" @default.
- W4315589177 modified "2023-09-26" @default.
- W4315589177 title "Unifying Nesterov's Accelerated Gradient Methods for Convex and Strongly Convex Objective Functions: From Continuous-Time Dynamics to Discrete-Time Algorithms" @default.
- W4315589177 doi "https://doi.org/10.48550/arxiv.2301.03576" @default.
- W4315589177 hasPublicationYear "2023" @default.
- W4315589177 type Work @default.
- W4315589177 citedByCount "0" @default.
- W4315589177 crossrefType "posted-content" @default.
- W4315589177 hasAuthorship W4315589177A5063507594 @default.
- W4315589177 hasAuthorship W4315589177A5075218521 @default.
- W4315589177 hasBestOaLocation W43155891771 @default.
- W4315589177 hasConcept C106159729 @default.
- W4315589177 hasConcept C112680207 @default.
- W4315589177 hasConcept C12108790 @default.
- W4315589177 hasConcept C126255220 @default.
- W4315589177 hasConcept C134306372 @default.
- W4315589177 hasConcept C145446738 @default.
- W4315589177 hasConcept C157972887 @default.
- W4315589177 hasConcept C162324750 @default.
- W4315589177 hasConcept C2524010 @default.
- W4315589177 hasConcept C28826006 @default.
- W4315589177 hasConcept C33923547 @default.
- W4315589177 hasConcept C34862557 @default.
- W4315589177 hasConcept C51544822 @default.
- W4315589177 hasConcept C72134830 @default.
- W4315589177 hasConcept C78045399 @default.
- W4315589177 hasConceptScore W4315589177C106159729 @default.
- W4315589177 hasConceptScore W4315589177C112680207 @default.
- W4315589177 hasConceptScore W4315589177C12108790 @default.
- W4315589177 hasConceptScore W4315589177C126255220 @default.
- W4315589177 hasConceptScore W4315589177C134306372 @default.
- W4315589177 hasConceptScore W4315589177C145446738 @default.
- W4315589177 hasConceptScore W4315589177C157972887 @default.
- W4315589177 hasConceptScore W4315589177C162324750 @default.
- W4315589177 hasConceptScore W4315589177C2524010 @default.
- W4315589177 hasConceptScore W4315589177C28826006 @default.
- W4315589177 hasConceptScore W4315589177C33923547 @default.
- W4315589177 hasConceptScore W4315589177C34862557 @default.
- W4315589177 hasConceptScore W4315589177C51544822 @default.
- W4315589177 hasConceptScore W4315589177C72134830 @default.
- W4315589177 hasConceptScore W4315589177C78045399 @default.
- W4315589177 hasLocation W43155891771 @default.
- W4315589177 hasOpenAccess W4315589177 @default.
- W4315589177 hasPrimaryLocation W43155891771 @default.
- W4315589177 hasRelatedWork W1914015964 @default.
- W4315589177 hasRelatedWork W1999577326 @default.
- W4315589177 hasRelatedWork W2001201947 @default.
- W4315589177 hasRelatedWork W2175532135 @default.
- W4315589177 hasRelatedWork W2356369451 @default.
- W4315589177 hasRelatedWork W2372442503 @default.
- W4315589177 hasRelatedWork W2394104306 @default.
- W4315589177 hasRelatedWork W3003165794 @default.
- W4315589177 hasRelatedWork W4287899549 @default.
- W4315589177 hasRelatedWork W783438975 @default.
- W4315589177 isParatext "false" @default.
- W4315589177 isRetracted "false" @default.
- W4315589177 workType "article" @default.