EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC-ND 4.0

Inverse Contrastive Representation Learning for Post-Click Conversion Rate Estimation

Ke Fei, Jingjing Li, Xinyue Zhang, Kun Zheng, Tianxu Ji, Wei Wang, Hongkai Qi

ACM Transactions on Information Systems · 30.06.2026 · 44 (5) · 1-26

DOI: 10.1145/3816726 ↗

Kurz gefasst

Fachartikel in der Zeitschrift ACM Transactions on Information Systems, online erschienen am 09.07.2026. Verfasst von Ke Fei und 6 weiteren Autor:innen. Die Arbeit nennt 46 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0). Gefördert von: National Natural Science Foundation of China, Fundamental Research Funds for the Central Universities, Science and Technology Plan Program of Huzhou City.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

In recommender systems, accurate post-click conversion rate (CVR) estimation is crucial for a company’s revenue growth. Currently, multi-task causal methods—such as multi-task doubly robust learning—improve CVR estimation by reducing bias and variance during training. However, these methods often overlook the impact of hard-to-classify samples, which deviate significantly from the sample cluster center and substantially increase the bias and variance of multi-task causal approaches. To address this limitation, we propose a novel inverse contrastive representation learning guided causal multi-task framework. Specifically, we first select clicked and unclicked samples that deviate significantly from their respective cluster centers to form negative pairs. We then apply inverse contrastive representation learning (ICRL) to pull these negative pairs closer together, encouraging deviated clicked samples to move closer to the clicked cluster center, and vice versa. This approach yields a more accurate and lower-variance inverse propensity score for the CVR estimator. Additionally, we compute the ICRL loss segmentally to reduce computational resource consumption. We conduct extensive experiments on six large-scale datasets, as well as online A/B tests on our advertising platform. Our method achieves a 2.9% improvement on offline datasets compared to state-of-the-art baselines, and a 0.6% increase over the online production model.

Förderung

  • National Natural Science Foundation of China
  • Fundamental Research Funds for the Central Universities
  • Science and Technology Plan Program of Huzhou City

Zitieren

APA

Fei, K., Li, J., Zhang, X., Zheng, K., Ji, T., Wang, W., & Qi, H. (2026). Inverse Contrastive Representation Learning for Post-Click Conversion Rate Estimation. ACM Transactions on Information Systems, 44(5), 1-26. https://doi.org/10.1145/3816726

Vancouver

Fei K, Li J, Zhang X, Zheng K, Ji T, Wang W, et al. Inverse Contrastive Representation Learning for Post-Click Conversion Rate Estimation. ACM Transactions on Information Systems. 2026;44(5):1-26. doi:10.1145/3816726

BibTeX
@article{Fei2026,
  author = {Fei, Ke and Li, Jingjing and Zhang, Xinyue and Zheng, Kun and Ji, Tianxu and Wang, Wei and Qi, Hongkai},
  title = {Inverse Contrastive Representation Learning for Post-Click Conversion Rate Estimation},
  journal = {ACM Transactions on Information Systems},
  year = {2026},
  volume = {44},
  number = {5},
  pages = {1--26},
  doi = {10.1145/3816726}
}
Kostenlos starten →