Multi-Faceted Large Embedding Tables for Pinterest Ads Ranking

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Su, Runze, Jin, Jiayin, Li, Jiacheng, Wang, Sihan, Bai, Guangtong, Wang, Zelun, Tang, Li, Meng, Yixiong, Wu, Huasen, Pan, Zhimeng, Li, Kungang, Sun, Han, Liu, Zhifang, Li, Haoyang, Ji, Siping, Peng, Degao, Zhuang, Jinfeng, Leng, Ling, Deshikachar, Prathibha
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866913985308131328
author Su, Runze
Jin, Jiayin
Li, Jiacheng
Wang, Sihan
Bai, Guangtong
Wang, Zelun
Tang, Li
Meng, Yixiong
Wu, Huasen
Pan, Zhimeng
Li, Kungang
Sun, Han
Liu, Zhifang
Li, Haoyang
Ji, Siping
Peng, Degao
Zhuang, Jinfeng
Leng, Ling
Deshikachar, Prathibha
author_facet Su, Runze
Jin, Jiayin
Li, Jiacheng
Wang, Sihan
Bai, Guangtong
Wang, Zelun
Tang, Li
Meng, Yixiong
Wu, Huasen
Pan, Zhimeng
Li, Kungang
Sun, Han
Liu, Zhifang
Li, Haoyang
Ji, Siping
Peng, Degao
Zhuang, Jinfeng
Leng, Ling
Deshikachar, Prathibha
contents Large embedding tables are indispensable in modern recommendation systems, thanks to their ability to effectively capture and memorize intricate details of interactions among diverse entities. As we explore integrating large embedding tables into Pinterest's ads ranking models, we encountered not only common challenges such as sparsity and scalability, but also several obstacles unique to our context. Notably, our initial attempts to train large embedding tables from scratch resulted in neutral metrics. To tackle this, we introduced a novel multi-faceted pretraining scheme that incorporates multiple pretraining algorithms. This approach greatly enriched the embedding tables and resulted in significant performance improvements. As a result, the multi-faceted large embedding tables bring great performance gain on both the Click-Through Rate (CTR) and Conversion Rate (CVR) domains. Moreover, we designed a CPU-GPU hybrid serving infrastructure to overcome GPU memory limits and elevate the scalability. This framework has been deployed in the Pinterest Ads system and achieved 1.34% online CPC reduction and 2.60% CTR increase with neutral end-to-end latency change.
format Preprint
id arxiv_https___arxiv_org_abs_2508_05700
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Multi-Faceted Large Embedding Tables for Pinterest Ads Ranking
Su, Runze
Jin, Jiayin
Li, Jiacheng
Wang, Sihan
Bai, Guangtong
Wang, Zelun
Tang, Li
Meng, Yixiong
Wu, Huasen
Pan, Zhimeng
Li, Kungang
Sun, Han
Liu, Zhifang
Li, Haoyang
Ji, Siping
Peng, Degao
Zhuang, Jinfeng
Leng, Ling
Deshikachar, Prathibha
Information Retrieval
Artificial Intelligence
Machine Learning
Large embedding tables are indispensable in modern recommendation systems, thanks to their ability to effectively capture and memorize intricate details of interactions among diverse entities. As we explore integrating large embedding tables into Pinterest's ads ranking models, we encountered not only common challenges such as sparsity and scalability, but also several obstacles unique to our context. Notably, our initial attempts to train large embedding tables from scratch resulted in neutral metrics. To tackle this, we introduced a novel multi-faceted pretraining scheme that incorporates multiple pretraining algorithms. This approach greatly enriched the embedding tables and resulted in significant performance improvements. As a result, the multi-faceted large embedding tables bring great performance gain on both the Click-Through Rate (CTR) and Conversion Rate (CVR) domains. Moreover, we designed a CPU-GPU hybrid serving infrastructure to overcome GPU memory limits and elevate the scalability. This framework has been deployed in the Pinterest Ads system and achieved 1.34% online CPC reduction and 2.60% CTR increase with neutral end-to-end latency change.
title Multi-Faceted Large Embedding Tables for Pinterest Ads Ranking
topic Information Retrieval
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2508.05700