mlsanityData to 7 Oct 2026

Balancing Unobserved Confounding with a Few Unbiased Ratings in Debiased Recommendations

arXiv 2304.09085 ↗First posted 17 Apr 2023cs.IR cs.LGcited by 18

Not on the map. Why?

Abstract (arXiv)

Recommender systems are seen as an effective tool to address information overload, but it is widely known that the presence of various biases makes direct training on large-scale observational data result in sub-optimal prediction performance. In contrast, unbiased ratings obtained from randomized controlled trials or A/B tests are considered to be the golden standard, but are costly and small in scale in reality. To exploit both types of data, recent works proposed to use unbiased ratings to correct the parameters of the propensity or imputation models trained on the biased dataset. However, the existing methods fail to obtain accurate predictions in the presence of unobserved confounding or model misspecification. In this paper, we propose a theoretically guaranteed model-agnostic balancing approach that can be applied to any existing debiasing method with the aim of combating unobserved confounding and model misspecification. The proposed approach makes full use of unbiased data by alternatively correcting model parameters learned with biased data, and adaptively learning balance coefficients of biased samples for further debiasing. Extensive real-world experiments are conducted along with the deployment of our proposal on four representative debiasing methods to demonstrate the effectiveness.

Key facts. Cited by 18 papers in the map (data through 7 Oct 2026). Builds on 26 of its 46 references that are in the map. First posted on arXiv 17 Apr 2023.

Builds on

26 of 46 references.

References · 26

  1. 2017Neural Collaborative Filteringcited by 1,108
  2. 2016Wide & Deep Learning for Recommender Systemscited by 745
  3. 2017DeepFM: A Factorization-Machine based Neural Network for CTR Predictioncited by 673
  4. 2018Learning to Reweight Examples for Robust Deep Learningcited by 670
  5. 2017Deep & Cross Network for Ad Click Predictionscited by 392
  6. 2016Recommendations as Treatments: Debiasing Learning and Evaluationcited by 246
  7. 2020Bias and Debias in Recommender System: A Survey and Future Directionscited by 239
  8. 2018Entire Space Multi-Task Model: An Effective Approach for Estimating Post-Click Conversion Ratecited by 172
  9. 2021Causal Intervention for Leveraging Popularity Bias in Recommendationcited by 116
  10. 2018Offline A/B testing for Recommender Systemscited by 94
  11. 2017Causal Embeddings for Recommendationcited by 79
  12. 2019Unbiased Recommender Learning from Missing-Not-At-Random Implicit Feedbackcited by 79
14 more
  1. 2018Confounding-Robust Policy Improvementcited by 73
  2. 2015Modeling User Exposure in Recommendationcited by 69
  3. 2021AutoDebias: Learning to Debias for Recommendationcited by 68
  4. 2018Removing Hidden Confounding by Experimental Groundingcited by 64
  5. 2021Deconfounded Recommendation for Alleviating Bias Amplificationcited by 51
  6. 2022ESCM2: Entire Space Counterfactual Multi-Task Model for Post-Click Conversion Rate Estimationcited by 47
  7. 2019Large-scale Causal Approaches to Debiasing Post-click Conversion Rate Estimation with Multi-task Learningcited by 39
  8. 2021Enhanced Doubly Robust Learning for Debiasing Post-click Conversion Rate Estimationcited by 28
  9. 2022On the Opportunity of Causal Learning in Recommendation Systems: Foundation, Estimation, Prediction and Challengescited by 28
  10. 2022StableDR: Stabilized Doubly Robust Learning for Recommendation on Data Missing Not at Randomcited by 24
  11. 2022TDR-CL: Targeted Doubly Robust Collaborative Learning for Debiased Recommendationscited by 18
  12. 2022A Generalized Doubly Robust Learning Framework for Debiasing Post-Click Conversion Rate Predictioncited by 18
  13. 2022Multiple Robust Learning for Recommendationcited by 11
  14. 2020Improving efficiency of inference in clinical trials with external control datacited by 10

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 3; 2024: 8; 2025: 5; 2026: 2.

18 papers.

Sort:

  1. 2023Optimal Transport for Treatment Effect Estimationcited by 32
  2. 2024How Do Recommendation Models Amplify Popularity Bias? An Analysis from the Spectral Perspectivecited by 21
  3. 2024Debiased Recommendation with Noisy Feedbackcited by 8
  4. 2024Doubly Calibrated Estimator for Recommendation on Data Missing Not At Randomcited by 7
  5. 2024Proximity Matters: Local Proximity Enhanced Balancing for Treatment Effect Estimationcited by 7
  6. 2024Debiased Collaborative Filtering with Kernel-Based Causal Balancingcited by 6
  7. 2024MetaCoCo: A New Few-Shot Classification Benchmark with Spurious Correlationcited by 5
  8. 2023ConvFormer: Revisiting Transformer for Sequential User Modelingcited by 3
  9. 2023Pareto Invariant Representation Learning for Multimedia Recommendationcited by 3
  10. 2024Entire-Space Variational Information Exploitation for Post-Click Conversion Rate Predictioncited by 3
  11. 2024Be Aware of the Neighborhood Effect: Modeling Selection Bias under Interferencecited by 2
  12. 2025Bi-Level Decision-Focused Causal Learning for Large-Scale Marketing Optimization: Bridging Observational and Experimental Datacited by 2
6 more
  1. 2025Optimal Policy Adaptation under Covariate Shiftcited by 0
  2. 2025A Partial Initialization Strategy to Mitigate the Overfitting Problem in CATE Estimation with Hidden Confoundingcited by 0
  3. 2025Addressing Correlated Latent Exogenous Variables in Debiased Recommender Systemscited by 0
  4. 2025BPL: Bias-adaptive Preference Distillation Learning for Recommender Systemcited by 0
  5. 2026PUID: A Personalized Deconfounding Framework for Recommender Systems under Hidden Confoundingcited by 0
  6. 2026The Pitfall of Scaling Up: Uncovering and Mitigating Popularity Bias Amplification in Scaling Transformer-based Recommenderscited by 0