mlsanityData to 14 Oct 2026

Interpretable Personalized Experimentation

arXiv 2111.03267 ↗First posted 5 Nov 2021cs.LG stat.MLcited by 3

Not on the map. Why?

Abstract (arXiv)

Black-box heterogeneous treatment effect (HTE) models are increasingly being used to create personalized policies that assign individuals to their optimal treatments. However, they are difficult to understand, and can be burdensome to maintain in a production environment. In this paper, we present a scalable, interpretable personalized experimentation system, implemented and deployed in production at Meta. The system works in a multiple treatment, multiple outcome setting typical at Meta to: (1) learn explanations for black-box HTE models; (2) generate interpretable personalized policies. We evaluate the methods used in the system on publicly available data and Meta use cases, and discuss lessons learnt during the development of the system.

Key facts. Cited by 3 papers in the map (data through 14 Oct 2026). Builds on 21 of its 60 references that are in the map. First posted on arXiv 5 Nov 2021.

Builds on

21 of 60 references.

References · 21

  1. 2015Distilling the Knowledge in a Neural Networkcited by 10,324
  2. 2015Estimation and Inference of Heterogeneous Treatment Effects using Random Forestscited by 600
  3. 2019Introduction to Multi-Armed Banditscited by 481
  4. 2017Meta-learners for Estimating Heterogeneous Treatment Effects using Machine Learningcited by 478
  5. 2017Quasi-Oracle Estimation of Heterogeneous Treatment Effectscited by 332
  6. 2016Learning Representations for Counterfactual Inferencecited by 329
  7. 2015Recursive Partitioning for Heterogeneous Causal Effectscited by 255
  8. 2018Offline Multi-Action Policy Learning: Generalization and Optimizationcited by 65
  9. 2018Policy Evaluation and Optimization with Continuous Treatmentscited by 58
  10. 2015Counterfactual Risk Minimization: Learning from Logged Bandit Feedbackcited by 33
  11. 2016Recursive Partitioning for Personalization using Observational Datacited by 28
  12. 2020Model Distillation for Revenue Optimization: Interpretable Personalized Pricingcited by 24
9 more
  1. 2019Bayesian Optimization for Policy Search via Online-Offline Experimentationcited by 20
  2. 2020Optimal Policy Treescited by 19
  3. 2020Stable discovery of interpretable subgroups via calibration in causal studiescited by 18
  4. 2021Value Function is All You Need: A Unified Learning Framework for Ride Hailing Platformscited by 16
  5. 2017Causal Rule Sets for Identifying Subgroups with Enhanced Treatment Effectcited by 15
  6. 2019Interpretable Subgroup Discovery in Treatment Effect Estimation with Application to Opioid Prescribing Guidelinescited by 13
  7. 2020Robust Recursive Partitioning for Heterogeneous Treatment Effects with Uncertainty Quantificationcited by 12
  8. 2013Personalized News Recommendation with Context Treescited by 11
  9. 2021Learning Optimal Prescriptive Trees from Observational Datacited by 10

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2022: 1; 2023: 1; 2024: 0; 2025: 0; 2026: 1.

3 papers.

  1. 2022Identifying Heterogeneous Treatment Effects in Multiple Outcomes using Joint Confidence Intervalscited by 4
  2. 2023Recovering Sparse and Interpretable Subgroups with Heterogeneous Treatment Effects with Censored Time-to-Event Outcomescited by 1
  3. 2026Decoding ML Decision: An Agentic Reasoning Framework for Large-Scale Ranking Systemcited by 1