EST: Towards Efficient Scaling Laws in Click-Through Rate Prediction via Unified Modeling
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866911440291495936 |
|---|---|
| author | Liu, Mingyang Bai, Yong Chan, Zhangming Chen, Sishuo Sheng, Xiang-Rong Zhu, Han Xu, Jian Chen, Xinyang |
| author_facet | Liu, Mingyang Bai, Yong Chan, Zhangming Chen, Sishuo Sheng, Xiang-Rong Zhu, Han Xu, Jian Chen, Xinyang |
| contents | Efficiently scaling industrial Click-Through Rate (CTR) prediction has recently attracted significant research attention. Existing approaches typically employ early aggregation of user behaviors to maintain efficiency. However, such non-unified or partially unified modeling creates an information bottleneck by discarding fine-grained, token-level signals essential for unlocking scaling gains. In this work, we revisit the fundamental distinctions between CTR prediction and Large Language Models (LLMs), identifying two critical properties: the asymmetry in information density between behavioral and non-behavioral features, and the modality-specific priors of content-rich signals. Accordingly, we propose the Efficiently Scalable Transformer (EST), which achieves fully unified modeling by processing all raw inputs in a single sequence without lossy aggregation. EST integrates two modules: Lightweight Cross-Attention (LCA), which prunes redundant self-interactions to focus on high-impact cross-feature dependencies, and Content Sparse Attention (CSA), which utilizes content similarity to dynamically select high-signal behaviors. Extensive experiments show that EST exhibits a stable and efficient power-law scaling relationship, enabling predictable performance gains with model scale. Deployed on Taobao's display advertising platform, EST significantly outperforms production baselines, delivering a 3.27\% RPM (Revenue Per Mile) increase and a 1.22\% CTR lift, establishing a practical pathway for scalable industrial CTR prediction models. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_10811 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | EST: Towards Efficient Scaling Laws in Click-Through Rate Prediction via Unified Modeling Liu, Mingyang Bai, Yong Chan, Zhangming Chen, Sishuo Sheng, Xiang-Rong Zhu, Han Xu, Jian Chen, Xinyang Information Retrieval Efficiently scaling industrial Click-Through Rate (CTR) prediction has recently attracted significant research attention. Existing approaches typically employ early aggregation of user behaviors to maintain efficiency. However, such non-unified or partially unified modeling creates an information bottleneck by discarding fine-grained, token-level signals essential for unlocking scaling gains. In this work, we revisit the fundamental distinctions between CTR prediction and Large Language Models (LLMs), identifying two critical properties: the asymmetry in information density between behavioral and non-behavioral features, and the modality-specific priors of content-rich signals. Accordingly, we propose the Efficiently Scalable Transformer (EST), which achieves fully unified modeling by processing all raw inputs in a single sequence without lossy aggregation. EST integrates two modules: Lightweight Cross-Attention (LCA), which prunes redundant self-interactions to focus on high-impact cross-feature dependencies, and Content Sparse Attention (CSA), which utilizes content similarity to dynamically select high-signal behaviors. Extensive experiments show that EST exhibits a stable and efficient power-law scaling relationship, enabling predictable performance gains with model scale. Deployed on Taobao's display advertising platform, EST significantly outperforms production baselines, delivering a 3.27\% RPM (Revenue Per Mile) increase and a 1.22\% CTR lift, establishing a practical pathway for scalable industrial CTR prediction models. |
| title | EST: Towards Efficient Scaling Laws in Click-Through Rate Prediction via Unified Modeling |
| topic | Information Retrieval |
| url | https://arxiv.org/abs/2602.10811 |