Probabilistic Rank and Reward: A Scalable Model for Slate Recommendation
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2022
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866913416594063360 |
|---|---|
| author | Aouali, Imad Hammou, Achraf Ait Sidi Sakhi, Otmane Rohde, David Vasile, Flavian |
| author_facet | Aouali, Imad Hammou, Achraf Ait Sidi Sakhi, Otmane Rohde, David Vasile, Flavian |
| contents | We introduce Probabilistic Rank and Reward (PRR), a scalable probabilistic model for personalized slate recommendation. Our approach allows off-policy estimation of the reward in the scenario where the user interacts with at most one item from a slate of K items. We show that the probability of a slate being successful can be learned efficiently by combining the reward, whether the user successfully interacted with the slate, and the rank, the item that was selected within the slate. PRR outperforms existing off-policy reward optimizing methods and is far more scalable to large action spaces. Moreover, PRR allows fast delivery of recommendations powered by maximum inner product search (MIPS), making it suitable in low latency domains such as computational advertising. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2208_06263 |
| institution | arXiv |
| publishDate | 2022 |
| record_format | arxiv |
| spellingShingle | Probabilistic Rank and Reward: A Scalable Model for Slate Recommendation Aouali, Imad Hammou, Achraf Ait Sidi Sakhi, Otmane Rohde, David Vasile, Flavian Information Retrieval Machine Learning We introduce Probabilistic Rank and Reward (PRR), a scalable probabilistic model for personalized slate recommendation. Our approach allows off-policy estimation of the reward in the scenario where the user interacts with at most one item from a slate of K items. We show that the probability of a slate being successful can be learned efficiently by combining the reward, whether the user successfully interacted with the slate, and the rank, the item that was selected within the slate. PRR outperforms existing off-policy reward optimizing methods and is far more scalable to large action spaces. Moreover, PRR allows fast delivery of recommendations powered by maximum inner product search (MIPS), making it suitable in low latency domains such as computational advertising. |
| title | Probabilistic Rank and Reward: A Scalable Model for Slate Recommendation |
| topic | Information Retrieval Machine Learning |
| url | https://arxiv.org/abs/2208.06263 |