Pure Exploration in Asynchronous Federated Bandits
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2023
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866916412942974976 |
|---|---|
| author | Wang, Zichen Li, Chuanhao Song, Chenyu Wang, Lianghui Gu, Quanquan Wang, Huazheng |
| author_facet | Wang, Zichen Li, Chuanhao Song, Chenyu Wang, Lianghui Gu, Quanquan Wang, Huazheng |
| contents | We study the federated pure exploration problem of multi-armed bandits and linear bandits, where $M$ agents cooperatively identify the best arm via communicating with the central server. To enhance the robustness against latency and unavailability of agents that are common in practice, we propose the first federated asynchronous multi-armed bandit and linear bandit algorithms for pure exploration with fixed confidence. Our theoretical analysis shows the proposed algorithms achieve near-optimal sample complexities and efficient communication costs in a fully asynchronous environment. Moreover, experimental results based on synthetic and real-world data empirically elucidate the effectiveness and communication cost-efficiency of the proposed algorithms. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2310_11015 |
| institution | arXiv |
| publishDate | 2023 |
| record_format | arxiv |
| spellingShingle | Pure Exploration in Asynchronous Federated Bandits Wang, Zichen Li, Chuanhao Song, Chenyu Wang, Lianghui Gu, Quanquan Wang, Huazheng Machine Learning We study the federated pure exploration problem of multi-armed bandits and linear bandits, where $M$ agents cooperatively identify the best arm via communicating with the central server. To enhance the robustness against latency and unavailability of agents that are common in practice, we propose the first federated asynchronous multi-armed bandit and linear bandit algorithms for pure exploration with fixed confidence. Our theoretical analysis shows the proposed algorithms achieve near-optimal sample complexities and efficient communication costs in a fully asynchronous environment. Moreover, experimental results based on synthetic and real-world data empirically elucidate the effectiveness and communication cost-efficiency of the proposed algorithms. |
| title | Pure Exploration in Asynchronous Federated Bandits |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2310.11015 |