Research map / Probabilistic and scientific modeling / Molecules and proteins
Doubly Robust Joint Learning for Recommendation on Data Missing Not at Random
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 imputed errors and learned propensities (MF-DR-JL vs MF-IPS) | MF-IPS · COAT MAR | MAE: 0.860 -> 0.778 | Helped |
| add learned propensities to MF-JL (MF-DR-JL vs MF-JL) (ablation) | MF-JL · COAT MAR | MAE: 0.866 -> 0.778 | Helped |
| add imputed errors to MF-IPS (MF-DR-JL vs MF-IPS) (ablation) | MF-IPS · YAHOO MAR | MSE: 0.989 -> 0.966 | Helped |
| add learned propensities to MF-JL (ablation) | MF-JL · YAHOO MAR | MAE: 0.899 -> 0.747 | Helped |
| heuristic imputation MF-HI vs plain MF | MF · COAT MAR | MAE: 0.920 -> 0.922 | Hurt |
| MF-MNAR generative assumptions vs MF | MF · YAHOO MAR | MAE: 1.154 -> 1.177 | Hurt |
| use uniform propensities (MF-JL) vs learned (MF-DR-JL) (ablation) | MF-DR-JL · AMAZON MNAR | MSE: 0.868 -> 0.871 | Mixed |
| learned propensities evaluated with MSE-SNIPS (ablation) | MF-JL · AMAZON MNAR | MSE-SNIPS: 0.851 -> 0.844 | Mixed |
| replace MF with NFM backbone in DR-JL | MF-DR-JL · COAT MAR | MAE: 0.778 -> 0.756 | Helped |
| DR-JL instead of IPS with NFM | NFM-IPS · COAT MAR | MAE: 0.832 -> 0.756 | Helped |
| DR estimator instead of IPS for bias | IPS · Simulated YAHOO FOUR | bias % under MSE: 66.8 -> 24.1 | Helped |
| DR estimator instead of IPS for bias | IPS · Simulated YAHOO ONE | bias % under MSE: 20.7 -> 9.9 | Helped |
Closest papers by meaning
- Multiple Robust Learning for Recommendation (AAAI 2023 · similarity 0.91)
- Relaxing the Accurate Imputation Assumption in Doubly Robust Learning for Debiased Collaborative Filtering (ICML 2024 · similarity 0.89)
- Fine-Grained Dynamic Framework for Bias-Variance Joint Optimization on Data Missing Not at Random (NeurIPS 2024 · similarity 0.86)
- Addressing Hidden Confounding with Heterogeneous Observational Datasets for Recommendation (NeurIPS 2024 · similarity 0.84)
- From Pairwise to Ranking: Climbing the Ladder to Ideal Collaborative Filtering with Pseudo-Ranking (AAAI 2025 · similarity 0.83)
- Both Supply and Precision: Sample Debias and Ranking Consistency Joint Learning for Large Scale Pre-Ranking System (AAAI 2025 · similarity 0.83)
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.