Tighter Generalisation Bounds via Interpolation
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2024
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866913226210410496 |
|---|---|
| author | Viallard, Paul Haddouche, Maxime Şimşekli, Umut Guedj, Benjamin |
| author_facet | Viallard, Paul Haddouche, Maxime Şimşekli, Umut Guedj, Benjamin |
| contents | This paper contains a recipe for deriving new PAC-Bayes generalisation bounds based on the $(f, Γ)$-divergence, and, in addition, presents PAC-Bayes generalisation bounds where we interpolate between a series of probability divergences (including but not limited to KL, Wasserstein, and total variation), making the best out of many worlds depending on the posterior distributions properties. We explore the tightness of these bounds and connect them to earlier results from statistical learning, which are specific cases. We also instantiate our bounds as training objectives, yielding non-trivial guarantees and practical performances. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2402_05101 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Tighter Generalisation Bounds via Interpolation Viallard, Paul Haddouche, Maxime Şimşekli, Umut Guedj, Benjamin Machine Learning This paper contains a recipe for deriving new PAC-Bayes generalisation bounds based on the $(f, Γ)$-divergence, and, in addition, presents PAC-Bayes generalisation bounds where we interpolate between a series of probability divergences (including but not limited to KL, Wasserstein, and total variation), making the best out of many worlds depending on the posterior distributions properties. We explore the tightness of these bounds and connect them to earlier results from statistical learning, which are specific cases. We also instantiate our bounds as training objectives, yielding non-trivial guarantees and practical performances. |
| title | Tighter Generalisation Bounds via Interpolation |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2402.05101 |