mlsanityData to 14 Oct 2026

Deconfounded Warm-Start Thompson Sampling with Applications to Precision Medicine

arXiv 2505.17283 ↗First posted 22 May 2025stat.ML cs.LG math.OCcited by 0

Not on the map. Why?

Abstract (arXiv)

Randomized clinical trials often require large patient cohorts before drawing definitive conclusions, yet abundant observational data from parallel studies remains underutilized due to confounding and hidden biases. To bridge this gap, we propose Deconfounded Warm-Start Thompson Sampling (DWTS), a practical approach that leverages a Doubly Debiased LASSO (DDL) procedure to identify a sparse set of reliable measured covariates and combines them with key hidden covariates to form a reduced context. By initializing Thompson Sampling (LinTS) priors with DDL-estimated means and variances on these measured features -- while keeping uninformative priors on hidden features -- DWTS effectively harnesses confounded observational data to kick-start adaptive clinical trials. Evaluated on both a purely synthetic environment and a virtual environment created using real cardiovascular risk dataset, DWTS consistently achieves lower cumulative regret than standard LinTS, showing how offline causal insights from observational data can improve trial efficiency and support more personalized treatment decisions.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 23 of its 47 references that are in the map. First posted on arXiv 22 May 2025.

Builds on

23 of 47 references.

References · 23

  1. 2015Estimation and Inference of Heterogeneous Treatment Effects using Random Forestscited by 600
  2. 2017Meta-learners for Estimating Heterogeneous Treatment Effects using Machine Learningcited by 478
  3. 2017Causal Effect Inference with Deep Latent-Variable Modelscited by 410
  4. 2019Adapting Neural Networks for the Estimation of Treatment Effectscited by 287
  5. 2021Nonparametric Estimation of Heterogeneous Treatment Effects: From Theory to Learning Algorithmscited by 144
  6. 2018The Blessings of Multiple Causescited by 104
  7. 2018Removing Hidden Confounding by Experimental Groundingcited by 64
  8. 2022Leveraging Offline Data in Online Reinforcement Learningcited by 42
  9. 2022Combining Observational and Randomized Data for Estimating Heterogeneous Treatment Effectscited by 32
  10. 2018Combining multiple observational data sources to estimate causal effectscited by 29
  11. 2020Combining Observational and Experimental Datasets Using Shrinkage Estimatorscited by 28
  12. 2021Estimating Average Treatment Effects via Orthogonal Regularizationcited by 25
11 more
  1. 2020Bandits with Partially Observable Confounded Datacited by 22
  2. 2021Adaptive Combination of Randomized and Observational Datacited by 22
  3. 2019Warm-starting Contextual Bandits: Robustly Combining Supervised and Bandit Feedbackcited by 21
  4. 2021Deconfounding Temporal Autoencoder: Estimating Treatment Effects over Time Using Noisy Proxiescited by 16
  5. 2022Artificial Replay: A Meta-Algorithm for Harnessing Historical Data in Banditscited by 13
  6. 2023Leveraging Demonstrations to Improve Online Learning: Quality Matterscited by 13
  7. 2020Estimating heterogeneous survival treatment effect in observational data using machine learningcited by 12
  8. 2020Doubly Debiased Lasso: High-Dimensional Inference under Hidden Confoundingcited by 9
  9. 2025Linear Bandits with Partially Observable Featurescited by 4
  10. 2023Efficient Online Learning with Offline Datasets for Infinite Horizon MDPs: A Bayesian Approachcited by 3
  11. 2022On LASSO for High Dimensional Predictive Regressioncited by 1

Built on by

Not cited yet.