EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Rethinking Homogeneity in Interactive Recommendation: Dynamic Fine-Grained Control for Mitigating Filter Bubbles

Pu Li, Xiaoyu Shi, Hong Xie, Chongjun Xia, Mingsheng Shang

ACM Transactions on the Web · 31.08.2026 · 20 (3) · 1-35

DOI: 10.1145/3828182 ↗

Kurz gefasst

Fachartikel in der Zeitschrift ACM Transactions on the Web, online erschienen am 14.08.2026. Verfasst von Pu Li und 4 weiteren Autor:innen. Die Arbeit nennt 75 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Natural Science Foundation of China, Strategic Priority Research Program of Chinese Academy of Sciences, Science and Technology Innovation Key R&D Program of Chongqing.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Recommender systems (RSs) are widely deployed in domains such as e-commerce and online media, yet often intensify the filter bubble problem by repeatedly exposing similar items. While prior studies generally regard such homogeneity as detrimental, its nuanced effects in interactive recommendation systems (IRSs) remain underexplored. In particular, the dynamic feedback loop between user preferences and system responses complicates the relationship between homogeneity, diversity, and engagement. This article presents the first fine-grained analysis of homogeneity in IRS, revealing that homogeneous recommendations can be beneficial in the early repetition stages before becoming detrimental. Building on this insight, we propose Diversity- and Debias-aware Interactive Recommendation (DDIR), an offline reinforcement learning framework that dynamically balances homogeneity and diversity over user sessions. DDIR employs a Transformer-based state encoder to capture time-aware diversity preferences, a homogeneity-friendly diversity model to estimate evolving tolerance to repetition, and a debiased interest model that disentangles intrinsic preferences from conformity bias while considering item quality. These components jointly guide a policy learner to sustain long-term engagement. Experiments on two realistic RL environments, KuaiEnv and KuaiRand, constructed from large-scale short-video logs, demonstrate that DDIR effectively mitigates filter bubbles, increasing average session length by 4.35% and long-term engagement by 25.65% over state-of-the-art baselines in KuaiRand simulated environment. Codes are available at https://github.com/16061025/DDIR .

Förderung

  • National Natural Science Foundation of China
  • Strategic Priority Research Program of Chinese Academy of Sciences
  • Science and Technology Innovation Key R&D Program of Chongqing
  • Natural Science Foundation of Chongqing

Zitieren

APA

Li, P., Shi, X., Xie, H., Xia, C., & Shang, M. (2026). Rethinking Homogeneity in Interactive Recommendation: Dynamic Fine-Grained Control for Mitigating Filter Bubbles. ACM Transactions on the Web, 20(3), 1-35. https://doi.org/10.1145/3828182

Vancouver

Li P, Shi X, Xie H, Xia C, Shang M. Rethinking Homogeneity in Interactive Recommendation: Dynamic Fine-Grained Control for Mitigating Filter Bubbles. ACM Transactions on the Web. 2026;20(3):1-35. doi:10.1145/3828182

BibTeX
@article{Li2026,
  author = {Li, Pu and Shi, Xiaoyu and Xie, Hong and Xia, Chongjun and Shang, Mingsheng},
  title = {Rethinking Homogeneity in Interactive Recommendation: Dynamic Fine-Grained Control for Mitigating Filter Bubbles},
  journal = {ACM Transactions on the Web},
  year = {2026},
  volume = {20},
  number = {3},
  pages = {1--35},
  doi = {10.1145/3828182}
}
Kostenlos starten →