Adaptive first-order methods with enhanced worst-case rates
Fuente:
arXiv
Salvato in:
| Autore principale: | |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866909201019699200 |
|---|---|
| author | Florea, Mihai I. |
| author_facet | Florea, Mihai I. |
| contents | The Optimized Gradient Method (OGM), its strongly convex extension, the Information Theoretical Exact Method (ITEM), as well as the related Triple Momentum Method (TMM) have superior convergence guarantees when compared to the Fast Gradient Method but lack adaptivity and their derivation is incompatible with composite problems. In this work we introduce a slightly modified version of the estimate sequence that can be used to simultaneously derive OGM, ITEM and TMM while adding memory along with the ability to dynamically adjust the convergence guarantees at runtime. Our framework can be extended to the composite setup and we use it to construct an Enhanced Accelerated Composite Gradient Method equipped with fully-adaptive line-search. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2405_07926 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Adaptive first-order methods with enhanced worst-case rates Florea, Mihai I. Optimization and Control 68Q25, 65Y20, 65B99, 90C25, 52A41 The Optimized Gradient Method (OGM), its strongly convex extension, the Information Theoretical Exact Method (ITEM), as well as the related Triple Momentum Method (TMM) have superior convergence guarantees when compared to the Fast Gradient Method but lack adaptivity and their derivation is incompatible with composite problems. In this work we introduce a slightly modified version of the estimate sequence that can be used to simultaneously derive OGM, ITEM and TMM while adding memory along with the ability to dynamically adjust the convergence guarantees at runtime. Our framework can be extended to the composite setup and we use it to construct an Enhanced Accelerated Composite Gradient Method equipped with fully-adaptive line-search. |
| title | Adaptive first-order methods with enhanced worst-case rates |
| topic | Optimization and Control 68Q25, 65Y20, 65B99, 90C25, 52A41 |
| url | https://arxiv.org/abs/2405.07926 |