BanditLP: Large-Scale Stochastic Optimization for Personalized Recommendations
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866911391780175872 |
|---|---|
| author | Nguyen, Phuc Zelditch, Benjamin Chen, Joyce Patra, Rohit Wei, Changshuai |
| author_facet | Nguyen, Phuc Zelditch, Benjamin Chen, Joyce Patra, Rohit Wei, Changshuai |
| contents | We present BanditLP, a scalable multi-stakeholder contextual bandit framework that unifies neural Thompson Sampling for learning objective-specific outcomes with a large-scale linear program for constrained action selection at serving time. The methodology is application-agnostic, compatible with arbitrary neural architectures, and deployable at web scale, with an LP solver capable of handling billions of variables. Experiments on public benchmarks and synthetic data show consistent gains over strong baselines. We apply this approach in LinkedIn's email marketing system and demonstrate business win, illustrating the value of integrated exploration and constrained optimization in production. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_15552 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | BanditLP: Large-Scale Stochastic Optimization for Personalized Recommendations Nguyen, Phuc Zelditch, Benjamin Chen, Joyce Patra, Rohit Wei, Changshuai Machine Learning Artificial Intelligence We present BanditLP, a scalable multi-stakeholder contextual bandit framework that unifies neural Thompson Sampling for learning objective-specific outcomes with a large-scale linear program for constrained action selection at serving time. The methodology is application-agnostic, compatible with arbitrary neural architectures, and deployable at web scale, with an LP solver capable of handling billions of variables. Experiments on public benchmarks and synthetic data show consistent gains over strong baselines. We apply this approach in LinkedIn's email marketing system and demonstrate business win, illustrating the value of integrated exploration and constrained optimization in production. |
| title | BanditLP: Large-Scale Stochastic Optimization for Personalized Recommendations |
| topic | Machine Learning Artificial Intelligence |
| url | https://arxiv.org/abs/2601.15552 |