AI Paper Atlas

Research map / Reinforcement learning / Policy gradient and actor-critic

Personalized Policy Learning through Discrete Experimentation

Zhiqi Zhang, Zhiyu Zeng, Ruohan Zhan, Dennis Zhang

ICML 2026 · Reinforcement learning › Policy gradient and actor-critic · approach: Confounding-robust policy evaluation

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use DLPT instead of pure deep learning (PDL) for ATE extrapolationstructured DNN · Held-out arm ATE recoveryMAPE (%): 7.18 -> 2.25Improved
use DLPT instead of linear regression for ATE extrapolationstructured DNN · Held-out arm ATE recoveryMAPE (%): 38.23 -> 2.25Improved
use DLPT instead of GRF T-learner for ATE extrapolationstructured DNN · Held-out arm ATE recoveryMAPE (%): 40.43 -> 2.25Improved
use DLPT instead of PDL for ATE extrapolationstructured DNN · Held-out arm ATE recoveryMSE: 32.82 -> 3.46Improved
use DLPT instead of PDL for policy learningstructured DNN · Policy learning, finite discrete classMPR (%): 6.32 -> 0.48Improved
use DLPT instead of best benchmark for optimal-arm accuracystructured DNN · Policy learning, finite discrete classAccuracy (%): 25.50 -> 73.08Improved
use DLPT instead of best benchmark for weighted accuracystructured DNN · Policy learning, finite discrete classWeighted accuracy (%): 33.79 -> 79.83Improved
use DLPT instead of PDL in finite observed policy classstructured DNN · Semi-synthetic, Finite Observed policy classMPR (%): 27.58 -> 17.96Improved
use DLPT instead of PDL in DNN policy classstructured DNN · Semi-synthetic, DNN policy classMPR (%): 7.42 -> 0.99Improved
use continuous DNN policy class instead of finite observed policy class for DLPTDLPT · Semi-syntheticMPR (%): 17.96 -> 0.99Improved
use DNN policy class instead of finite observed class for simple parametric estimators (LR, LogR, LCM)LR/LogR/LCM · Semi-syntheticMPR (%): 29.89 -> 12.65Hurt
use linear policy class instead of finite observed class for DLPTDLPT · Semi-syntheticMPR (%): 17.96 -> 0.01Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy