Doubly Robust

Doubly robust (DR) methods are statistical techniques designed to provide robust causal inference and unbiased predictions from observational data, mitigating biases arising from missing data or confounding variables. Current research focuses on extending DR estimators to handle various complexities, including noisy feedback, continuous treatments, and network effects, often integrating machine learning models like variational autoencoders and graph neural networks for improved accuracy and efficiency. The significance of DR lies in its ability to produce reliable causal estimates and predictions in diverse settings, impacting fields ranging from recommendation systems and causal inference to policy evaluation and treatment effect estimation.

Papers

November 12, 2024

Doubly Robust Regression Discontinuity Designs
Masahiro Kato
Treatment Effect Nonparametric Regression Doubly Robust Novel Estimator Discontinuity Discrimination

June 24, 2024

Debiased Recommendation with Noisy Feedback
Haoxuan Li, Chunyuan Zheng, Wenjie Wang, Hao Wang, Fuli Feng, Xiao-Hua Zhou
Recommender System Unbiased Learning Debiased Recommendation Noisy Feedback Doubly Robust Missing Not at Random

June 2, 2024

A Tutorial on Doubly Robust Learning for Causal Inference
Hlynur Davíð Hlynsson
Causal Inference Robust Learning Tutorial Review Multiple Outcome Propensity Score Doubly Robust

March 17, 2024

Graph Machine Learning based Doubly Robust Estimator for Network Causal Effects
Seyedeh Baharan Khatami, Harsh Parikh, Haowei Chen, Sudeepa Roy, Babak Salimi
Social Medium Causal Effect Graph Machine Learning Doubly Robust Peer Effect Network Effect

September 22, 2023

Doubly Robust Proximal Causal Learning for Continuous Treatments
Yong Wu, Yanwei Fu, Shouyan Wang, Xinwei Sun
Causal Effect Estimation Influence Function Nuisance Parameter Valued Treatment Doubly Robust Proximal Causal

August 13, 2023

CDR: Conservative Doubly Robust Learning for Debiased Recommendation
ZiJie Song, JiaWei Chen, Sheng Zhou, QiHao Shi, Yan Feng, Chun Chen, Can Wang
Recommendation System Robust Learning Knowledge Enhanced Conditional Imputation Debiased Recommendation Robust Strategy Doubly Robust

March 7, 2023

DR-VIDAL -- Doubly Robust Variational Information-theoretic Deep Adversarial Learning for Counterfactual Prediction and Treatment Effect Estimation on Real World Data
Shantanu Ghosh, Zheng Feng, Jiang Bian, Kevin Butler, Mattia Prosperi
Treatment Effect Real World Data Counterfactual Prediction Variational Objective Causal Assumption Doubly Robust

February 23, 2023

Asymptotically Unbiased Off-Policy Policy Evaluation when Reusing Old Data in Nonstationary Environments
Vincent Liu, Yash Chandak, Philip Thomas, Martha White
Contextual Bandit Policy Evaluation Finite Horizon Data Recycling Doubly Robust Nonstationary Environment

January 15, 2023

Doubly Robust Counterfactual Classification
Kwangho Kim, Edward H. Kennedy, José R. Zubizarreta
High Quality Counterfactuals Doubly Robust Recidivism Model Counterfactual Classification

December 9, 2022

Doubly Robust Kernel Statistics for Testing Distributional Treatment Effects
Jake Fawkes, Robert Hu, Robin J. Evans, Dino Sejdinovic
Causal Inference Treatment Effect Doubly Robust Mean Embeddings

November 12, 2022

A Generalized Doubly Robust Learning Framework for Debiasing Post-Click Conversion Rate Prediction
Quanyu Dai, Haoxuan Li, Peng Wu, Zhenhua Dong, Xiao-Hua Zhou, Rui Zhang, Rui zhang, Jie Sun
Generalization Performance Generalization Bound Doubly Robust Conversion Rate Prediction Robustness Bias Post Click

March 31, 2022

Doubly-Robust Estimation for Correcting Position-Bias in Click Feedback for Unbiased Learning to Rank
Harrie Oosterhuis
Stable Rank Unbiased Learning Learning to Rank Doubly Robust Unbiased Learning to Rank Click Feedback Robust Counterfactuals

December 18, 2021

Off-Policy Evaluation Using Information Borrowing and Context-Based Switching
Sutanoy Dasgupta, Yabo Niu, Kishan Panaganti, Dileep Kalathil, Debdeep Pati, Bani Mallick
Contextual Bandit Policy Evaluation Inverse Propensity Information Acquisition Doubly Robust