Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Generative Recommendation for Large-Scale Advertising

Ben Xue, Dan Liu, Lixiang Wang, Mingjie Sun, Peng Wang, Pengfei Zhang, Shaoyun Shi, Tianyu Xu, Yunhao Sha, Zhiqiang Liu, Bo Kong, Bo Wang, Hang Yang, Jieting Xue, Junhao Wang, Shengyu Wang, Shuping Hui, Wencai Ye, Xiao Lin, Yongzhi Li, Yuhang Chen, Zhihui Yin, Quan Chen, Shiyang Wen, Wenjin Wu, Han Li, Guorui Zhou, Changcheng Li, Peng Jiang

TL;DR

This work presents a production-oriented generative recommender co-designed across architecture, learning, and serving, named GR4AD (Generative Recommendation for ADdvertising), which has been fully deployed in Kuaishou advertising system with over 400 million users and achieves high-throughput real-time serving.

Abstract

Generative recommendation has recently attracted widespread attention in industry due to its potential for scaling and stronger model capacity. However, deploying real-time generative recommendation in large-scale advertising requires designs beyond large-language-model (LLM)-style training and serving recipes. We present a production-oriented generative recommender co-designed across architecture, learning, and serving, named GR4AD (Generative Recommendation for ADdvertising). As for tokenization, GR4AD proposes UA-SID (Unified Advertisement Semantic ID) to capture complicated business information. Furthermore, GR4AD introduces LazyAR, a lazy autoregressive decoder that relaxes layer-wise dependencies for short, multi-candidate generation, preserving effectiveness while reducing inference cost, which facilitates scaling under fixed serving budgets. To align optimization with business value, GR4AD employs VSL (Value-Aware Supervised Learning) and proposes RSPO (Ranking-Guided Softmax Preference Optimization), a ranking-aware, list-wise reinforcement learning algorithm that optimizes value-based rewards under list-level metrics for continual online updates. For online inference, we further propose dynamic beam serving, which adapts beam width across generation levels and online load to control compute. Large-scale online A/B tests show up to 4.2% ad revenue improvement over an existing DLRM-based stack, with consistent gains from both model scaling and inference-time scaling. GR4AD has been fully deployed in Kuaishou advertising system with over 400 million users and achieves high-throughput real-time serving.

Generative Recommendation for Large-Scale Advertising

TL;DR

This work presents a production-oriented generative recommender co-designed across architecture, learning, and serving, named GR4AD (Generative Recommendation for ADdvertising), which has been fully deployed in Kuaishou advertising system with over 400 million users and achieves high-throughput real-time serving.

Abstract

Generative recommendation has recently attracted widespread attention in industry due to its potential for scaling and stronger model capacity. However, deploying real-time generative recommendation in large-scale advertising requires designs beyond large-language-model (LLM)-style training and serving recipes. We present a production-oriented generative recommender co-designed across architecture, learning, and serving, named GR4AD (Generative Recommendation for ADdvertising). As for tokenization, GR4AD proposes UA-SID (Unified Advertisement Semantic ID) to capture complicated business information. Furthermore, GR4AD introduces LazyAR, a lazy autoregressive decoder that relaxes layer-wise dependencies for short, multi-candidate generation, preserving effectiveness while reducing inference cost, which facilitates scaling under fixed serving budgets. To align optimization with business value, GR4AD employs VSL (Value-Aware Supervised Learning) and proposes RSPO (Ranking-Guided Softmax Preference Optimization), a ranking-aware, list-wise reinforcement learning algorithm that optimizes value-based rewards under list-level metrics for continual online updates. For online inference, we further propose dynamic beam serving, which adapts beam width across generation levels and online load to control compute. Large-scale online A/B tests show up to 4.2% ad revenue improvement over an existing DLRM-based stack, with consistent gains from both model scaling and inference-time scaling. GR4AD has been fully deployed in Kuaishou advertising system with over 400 million users and achieves high-throughput real-time serving.
Paper Structure (67 sections, 1 theorem, 27 equations, 6 figures, 4 tables)

This paper contains 67 sections, 1 theorem, 27 equations, 6 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Related Works
  3. Generative Recommendation
  4. Reinforcement Learning
  5. Semantic IDs
  6. Methodology
  7. Unified Advertisement Semantic ID
  8. Unified Advertisement Embedding (UAE)
  9. Instruction Tuning (IT)
  10. Co-occurrence Learning (CL)
  11. Multi-Granularity-Multi-Resolution (MGMR) RQ-Kmeans
  12. Lazy Autoregressive Decoder
  13. Vanilla Autoregressive Decoding.
  14. LazyAR: Late-Inject Autoregression.
  15. Why LazyAR is faster.
  16. ...and 52 more sections

Key Result

theorem 1

Ignoring constant terms, $\mathcal{L}_\text{RSPO}$ optimizes an upper bound of $\mathrm{NDCGcost}$.

Figures (6)

  • Figure 1: Overview of our proposed GR4AD: model architecture and learning algorithm.
  • Figure 2: Illustration of Unified Advertisement Semantic ID.
  • Figure 3: Comparison of Vanilla AR, DeepSeek MTP, LazyAR.
  • Figure 4: System overview: training and serving of GR4AD.
  • Figure 5: Scaling laws of model size and beam width.
  • ...and 1 more figures

Theorems & Definitions (1)

  • theorem 1