Distributed Experimental Design Networks
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_ | 1866909067580014592 |
|---|---|
| author | Li, Yuanyuan Su, Lili Joe-Wong, Carlee Yeh, Edmund Ioannidis, Stratis |
| author_facet | Li, Yuanyuan Su, Lili Joe-Wong, Carlee Yeh, Edmund Ioannidis, Stratis |
| contents | As edge computing capabilities increase, model learning deployments in diverse edge environments have emerged. In experimental design networks, introduced recently, network routing and rate allocation are designed to aid the transfer of data from sensors to heterogeneous learners. We design efficient experimental design network algorithms that are (a) distributed and (b) use multicast transmissions. This setting poses significant challenges as classic decentralization approaches often operate on (strictly) concave objectives under differentiable constraints. In contrast, the problem we study here has a non-convex, continuous DR-submodular objective, while multicast transmissions naturally result in non-differentiable constraints. From a technical standpoint, we propose a distributed Frank-Wolfe and a distributed projected gradient ascent algorithm that, coupled with a relaxation of non-differentiable constraints, yield allocations within a $1-1/e$ factor from the optimal. Numerical evaluations show that our proposed algorithms outperform competitors with respect to model learning quality. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2401_04996 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Distributed Experimental Design Networks Li, Yuanyuan Su, Lili Joe-Wong, Carlee Yeh, Edmund Ioannidis, Stratis Networking and Internet Architecture As edge computing capabilities increase, model learning deployments in diverse edge environments have emerged. In experimental design networks, introduced recently, network routing and rate allocation are designed to aid the transfer of data from sensors to heterogeneous learners. We design efficient experimental design network algorithms that are (a) distributed and (b) use multicast transmissions. This setting poses significant challenges as classic decentralization approaches often operate on (strictly) concave objectives under differentiable constraints. In contrast, the problem we study here has a non-convex, continuous DR-submodular objective, while multicast transmissions naturally result in non-differentiable constraints. From a technical standpoint, we propose a distributed Frank-Wolfe and a distributed projected gradient ascent algorithm that, coupled with a relaxation of non-differentiable constraints, yield allocations within a $1-1/e$ factor from the optimal. Numerical evaluations show that our proposed algorithms outperform competitors with respect to model learning quality. |
| title | Distributed Experimental Design Networks |
| topic | Networking and Internet Architecture |
| url | https://arxiv.org/abs/2401.04996 |