LEADRE: Multi-Faceted Knowledge Enhanced LLM Empowered Display Advertisement Recommender System
Summary: LEADRE integrates LLMs into display-ad pipelines by fine-tuning with intent-aware prompt–response pairs and ad-specific alignment (auxiliary tasks + DPO) to leverage ad content and user intent beyond ID-based retrieval. Also uses hybrid, latency-aware deployment; offline gains and live A/B (+1.57%/+1.17% GMV on WeChat), serving tens of billions requests/day. (summarized by gpt-5-mini on Feb 09 2026)
Incoming Non-self Citations Over Time
No non-self incoming citations found for this paper in this database.
Authors
- 1. Fengxin Li
- 2. Yi Li
- 3. Yue Liu
- 4. Chao Zhou
- 5. Yuan Wang
- 6. Xiaoxiang Deng
- 7. Wei Xue
- 8. Dapeng Liu
- 9. Lei Xiao
- 10. Haijie Gu
- 11. Jie Jiang
- 12. Hongyan Liu
- 13. Biao Qin
- 14. Jun He
Incoming Citations (Sorted by Pagerank)
Showing 0 of 0 citing papers.
| Rank | Citing Paper | Year | Venue | Pagerank |
|---|
Previous
Page 1 / 1
Next
Outgoing Citations (Sorted by Pagerank)
Showing 0 of 0 cited papers.
Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.
| Rank | Cited Paper | Year | Venue | Pagerank |
|---|
Previous
Page 1 / 1
Next