mlsanityData to 7 Oct 2026

Enhancing Counterfactual Classification via Self-Training

arXiv 2112.04461 ↗First posted 8 Dec 2021cs.LG stat.MLcited by 5

Not on the map. Why?

Abstract (arXiv)

Unlike traditional supervised learning, in many settings only partial feedback is available. We may only observe outcomes for the chosen actions, but not the counterfactual outcomes associated with other alternatives. Such settings encompass a wide variety of applications including pricing, online marketing and precision medicine. A key challenge is that observational data are influenced by historical policies deployed in the system, yielding a biased data distribution. We approach this task as a domain adaptation problem and propose a self-training algorithm which imputes outcomes with categorical values for finite unseen actions in the observational data to simulate a randomized trial through pseudolabeling, which we refer to as Counterfactual Self-Training (CST). CST iteratively imputes pseudolabels and retrains the model. In addition, we show input consistency loss can further improve CST performance which is shown in recent theoretical analysis of pseudolabeling. We demonstrate the effectiveness of the proposed algorithms on both synthetic and real datasets.

Key facts. Cited by 5 papers in the map (data through 7 Oct 2026). Builds on 21 of its 56 references that are in the map. First posted on arXiv 8 Dec 2021.

Builds on

21 of 56 references.

References · 21

  1. 2014Adam: A Method for Stochastic Optimizationcited by 36,960
  2. 2017Virtual Adversarial Training: A Regularization Method for Supervised and Semi-Supervised Learningcited by 918
  3. 2017Meta-learners for Estimating Heterogeneous Treatment Effects using Machine Learningcited by 475
  4. 2019Confidence Regularized Self-Trainingcited by 418
  5. 2016Learning Representations for Counterfactual Inferencecited by 327
  6. 2021In Defense of Pseudo-Labeling: An Uncertainty-Aware Pseudo-label Selection Framework for Semi-Supervised Learningcited by 301
  7. 2019Adapting Neural Networks for the Estimation of Treatment Effectscited by 285
  8. 2020Theoretical Analysis of Self-Training with Deep Networks on Unlabeled Datacited by 179
  9. 2015Doubly Robust Policy Evaluation and Optimizationcited by 158
  10. 2017Bayesian Inference of Individualized Treatment Effects using Multi-task Gaussian Processescited by 156
  11. 2018Confounding-Robust Policy Improvementcited by 73
  12. 2020Off-policy Bandits with Deficient Supportcited by 60
9 more
  1. 2021Human-AI Collaboration with Bandit Feedbackcited by 36
  2. 2018Domain Adaptation for Semantic Segmentation via Class-Balanced Self-Trainingcited by 31
  3. 2016A Bayesian Nonparametric Approach for Estimating Individualized Treatment-Response Curvescited by 15
  4. 2019Unsupervised Domain Adaptation via Calibrating Uncertaintiescited by 15
  5. 2019Cost-Effective Incentive Allocation via Structured Counterfactual Inferencecited by 12
  6. 2016Large-scale Validation of Counterfactual Learning Methods: A Test-Bedcited by 8
  7. 2021Loss Functions for Discrete Contextual Pricing with Observational Datacited by 6
  8. 2019Classifying Treatment Responders Under Causal Effect Monotonicitycited by 4
  9. 2016The Power and Limits of Predictive Approaches to Observational-Data-Driven Optimizationcited by 3

Built on by

Cited by year

Citing papers per year: 2022: 2; 2023: 3.

5 papers.

  1. 2022Probabilistic Conformal Prediction Using Conditional Random Samplescited by 47
  2. 2023Learning Complementary Policies for Human-AI Teamscited by 7
  3. 2023Confounding-Robust Policy Improvement with Human-AI Teamscited by 4
  4. 2022Semi-supervised Batch Learning From Logged Datacited by 2
  5. 2023Multi-Action Dialog Policy Learning from Logged User Feedbackcited by 0