Research map / Probabilistic and scientific modeling / Treatment effects
Diversity Recommendation via Causal Deconfounding of Co-purchase Relations and Counterfactual Exposure
Jingmao Zhang, Zhiting Zhao, Yunqi Lin, Jianghong Ma, Tianjun Wei, Haijun Zhang, Xiaofeng Zhang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add counterfactual exposure and candidate selection (CSCE) (ablation) | Cadence · Toy | C@100: 35.4518 -> 64.7435 | Improved |
| add CSCE (ablation) | Cadence · Beauty | R@100: 0.2587 -> 0.2590 | No clear effect |
| add CSCE (ablation) | Cadence · TaoBao | C@100: 33.5962 -> 52.1118 | Improved |
| add CIGR (UACR-refined item graph) (ablation) | Cadence · TaoBao | R@100: 0.0683 -> 0.0699 | Improved |
| add CIGR (ablation) | Cadence · TaoBao | C@100: 49.2669 -> 52.1118 | Improved |
| add CIGR (ablation) | Cadence · Toy | C@100: 64.2619 -> 64.7435 | No clear effect |
| add both modules to base LightGCN (ablation) | Cadence · TaoBao | R@100: 0.0505 -> 0.0699 | Improved |
| plug Cadence into NCL | NCL · Toy | Coverage@100: 38.4338 -> 74.0611 | Improved |
| plug Cadence into DimCL | DimCL · Beauty | Recall@100: 0.2554 -> 0.2678 | Improved |
| plug Cadence into backbones on small dense dataset | DimCL/DCF/PPAC/NCL · Beauty | Recall@100: +2.70% to +4.86% recall vs +21.59% to +30.22% on Toy | Mixed |
| train LightGCN+BPR; embedding norm vs popularity | LightGCN · TaoBao | Pearson r of log norm vs log popularity: -0.041 -> 0.978 | Improved |
Closest papers by meaning
- Tree of Preferences for Diversified Recommendation (NeurIPS 2025 · similarity 0.90)
- Alleviating Dual Biases in Recommendation (Student Abstract) (AAAI 2025 · similarity 0.88)
- HERS: Modeling Influential Contexts with Heterogeneous Relations for Sparse and Cold-Start Recommendation (AAAI 2019 · similarity 0.87)
- Addressing Hidden Confounding with Heterogeneous Observational Datasets for Recommendation (NeurIPS 2024 · similarity 0.86)
- DivGCL: A Graph Contrastive Learning Model for Diverse Recommendation (AAAI 2025 · similarity 0.86)
- Robust Graph Based Social Recommendation Through Contrastive Multi-View Learning (AAAI 2025 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
