Enregistré dans:
| Auteurs principaux: | , , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | https://arxiv.org/abs/2603.04227 |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866911484601171968 |
|---|---|
| author | Li, Chenfei Zhao, Hantao Yao, Weixi Huang, Ruiming Lu, Rongrong Tian, Geng Kong, Dongying |
| author_facet | Li, Chenfei Zhao, Hantao Yao, Weixi Huang, Ruiming Lu, Rongrong Tian, Geng Kong, Dongying |
| contents | Optimizing reranking in advertising feeds is a constrained combinatorial problem, requiring simultaneous maximization of platform revenue and preservation of user experience. Recent generative ranking methods enable listwise optimization via autoregressive decoding, but their deployment is hindered by high inference latency and limited constraint handling.
We propose a constraint-aware generative reranking framework that transforms constrained optimization into bounded neural decoding. Unlike prior approaches that separate generator and evaluator models, our framework unifies sequence generation and reward estimation into a single network.
We further introduce constraint-aware reward pruning, integrating constraint satisfaction directly into decoding to efficiently generate optimal sequences. Experiments on large-scale industrial feeds and online A/B tests show that our method improves revenue and user engagement while meeting strict latency requirements, providing an efficient neural solution for constrained listwise optimization. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_04227 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Constraint-Aware Generative Re-ranking for Multi-Objective Optimization in Advertising Feeds Li, Chenfei Zhao, Hantao Yao, Weixi Huang, Ruiming Lu, Rongrong Tian, Geng Kong, Dongying Information Retrieval Optimizing reranking in advertising feeds is a constrained combinatorial problem, requiring simultaneous maximization of platform revenue and preservation of user experience. Recent generative ranking methods enable listwise optimization via autoregressive decoding, but their deployment is hindered by high inference latency and limited constraint handling. We propose a constraint-aware generative reranking framework that transforms constrained optimization into bounded neural decoding. Unlike prior approaches that separate generator and evaluator models, our framework unifies sequence generation and reward estimation into a single network. We further introduce constraint-aware reward pruning, integrating constraint satisfaction directly into decoding to efficiently generate optimal sequences. Experiments on large-scale industrial feeds and online A/B tests show that our method improves revenue and user engagement while meeting strict latency requirements, providing an efficient neural solution for constrained listwise optimization. |
| title | Constraint-Aware Generative Re-ranking for Multi-Objective Optimization in Advertising Feeds |
| topic | Information Retrieval |
| url | https://arxiv.org/abs/2603.04227 |