GRAB: An LLM-Inspired Sequence-First Click-Through Rate Prediction Modeling Paradigm
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866918320419110912 |
|---|---|
| author | Chen, Shaopeng Xie, Chuyue Ren, Huimin Zhang, Shaozong Zhang, Han Cheng, Ruobing Cao, Zhiqiang Ju, Zehao Gao, Yu Ding, Jie Chen, Xiaodong Jiao, Xuewu Li, Shuanglong Lin, Liu |
| author_facet | Chen, Shaopeng Xie, Chuyue Ren, Huimin Zhang, Shaozong Zhang, Han Cheng, Ruobing Cao, Zhiqiang Ju, Zehao Gao, Yu Ding, Jie Chen, Xiaodong Jiao, Xuewu Li, Shuanglong Lin, Liu |
| contents | Traditional Deep Learning Recommendation Models (DLRMs) face increasing bottlenecks in performance and efficiency, often struggling with generalization and long-sequence modeling. Inspired by the scaling success of Large Language Models (LLMs), we propose Generative Ranking for Ads at Baidu (GRAB), an end-to-end generative framework for Click-Through Rate (CTR) prediction. GRAB integrates a novel Causal Action-aware Multi-channel Attention (CamA) mechanism to effectively capture temporal dynamics and specific action signals within user behavior sequences. Full-scale online deployment demonstrates that GRAB significantly outperforms established DLRMs, delivering a 3.05% increase in revenue and a 3.49% rise in CTR. Furthermore, the model demonstrates desirable scaling behavior: its expressive power shows a monotonic and approximately linear improvement as longer interaction sequences are utilized. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_01865 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | GRAB: An LLM-Inspired Sequence-First Click-Through Rate Prediction Modeling Paradigm Chen, Shaopeng Xie, Chuyue Ren, Huimin Zhang, Shaozong Zhang, Han Cheng, Ruobing Cao, Zhiqiang Ju, Zehao Gao, Yu Ding, Jie Chen, Xiaodong Jiao, Xuewu Li, Shuanglong Lin, Liu Information Retrieval Artificial Intelligence Traditional Deep Learning Recommendation Models (DLRMs) face increasing bottlenecks in performance and efficiency, often struggling with generalization and long-sequence modeling. Inspired by the scaling success of Large Language Models (LLMs), we propose Generative Ranking for Ads at Baidu (GRAB), an end-to-end generative framework for Click-Through Rate (CTR) prediction. GRAB integrates a novel Causal Action-aware Multi-channel Attention (CamA) mechanism to effectively capture temporal dynamics and specific action signals within user behavior sequences. Full-scale online deployment demonstrates that GRAB significantly outperforms established DLRMs, delivering a 3.05% increase in revenue and a 3.49% rise in CTR. Furthermore, the model demonstrates desirable scaling behavior: its expressive power shows a monotonic and approximately linear improvement as longer interaction sequences are utilized. |
| title | GRAB: An LLM-Inspired Sequence-First Click-Through Rate Prediction Modeling Paradigm |
| topic | Information Retrieval Artificial Intelligence |
| url | https://arxiv.org/abs/2602.01865 |