Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2407.15439 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866910545359142912 |
|---|---|
| author | Chen, Ziqun Cai, Kechao Chen, Zhuoyue Zhang, Jinbei Lui, John C. S. |
| author_facet | Chen, Ziqun Cai, Kechao Chen, Zhuoyue Zhang, Jinbei Lui, John C. S. |
| contents | We study the stochastic combinatorial semi-bandit problem with unrestricted feedback delays under merit-based fairness constraints. This is motivated by applications such as crowdsourcing, and online advertising, where immediate feedback is not immediately available and fairness among different choices (or arms) is crucial. We consider two types of unrestricted feedback delays: reward-independent delays where the feedback delays are independent of the rewards, and reward-dependent delays where the feedback delays are correlated with the rewards. Furthermore, we introduce merit-based fairness constraints to ensure a fair selection of the arms. We define the reward regret and the fairness regret and present new bandit algorithms to select arms under unrestricted feedback delays based on their merits. We prove that our algorithms all achieve sublinear expected reward regret and expected fairness regret, with a dependence on the quantiles of the delay distribution. We also conduct extensive experiments using synthetic and real-world data and show that our algorithms can fairly select arms with different feedback delays. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2407_15439 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Merit-based Fair Combinatorial Semi-Bandit with Unrestricted Feedback Delays Chen, Ziqun Cai, Kechao Chen, Zhuoyue Zhang, Jinbei Lui, John C. S. Machine Learning We study the stochastic combinatorial semi-bandit problem with unrestricted feedback delays under merit-based fairness constraints. This is motivated by applications such as crowdsourcing, and online advertising, where immediate feedback is not immediately available and fairness among different choices (or arms) is crucial. We consider two types of unrestricted feedback delays: reward-independent delays where the feedback delays are independent of the rewards, and reward-dependent delays where the feedback delays are correlated with the rewards. Furthermore, we introduce merit-based fairness constraints to ensure a fair selection of the arms. We define the reward regret and the fairness regret and present new bandit algorithms to select arms under unrestricted feedback delays based on their merits. We prove that our algorithms all achieve sublinear expected reward regret and expected fairness regret, with a dependence on the quantiles of the delay distribution. We also conduct extensive experiments using synthetic and real-world data and show that our algorithms can fairly select arms with different feedback delays. |
| title | Merit-based Fair Combinatorial Semi-Bandit with Unrestricted Feedback Delays |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2407.15439 |