Lune

ICML2025Top-tier venue

Towards the Efficient Inference by Incorporating Automated Computational Phenotypes under Covariate Shift

Chao Ying, Jun Jin, Yi Guo, Xiudi Li, Muxuan Liang, Jiwei Zhao

Abstract

Collecting gold-standard phenotype data via manual extraction is typically labor-intensive and slow, whereas automated computational phenotypes (ACPs) offer a systematic and much faster alternative. However, simply replacing the goldstandard with ACPs, without acknowledging their differences, could lead to biased results and misleading conclusions. Motivated by the complexity of incorporating ACPs while maintaining the validity of downstream analyses, in this paper, we consider a semi-supervised learning setting that consists of both labeled data (with gold-standard) and unlabeled data (without gold-standard), under the covariate shift framework. We develop doubly robust and semiparametrically efficient estimators that leverage ACPs for general target parameters in the unlabeled and combined populations. In addition, we carefully analyze the efficiency gains achieved by incorporating ACPs, comparing scenarios with and without their inclusion. Notably, we identify that ACPs for the unlabeled data, instead of for the labeled data, drive the enhanced efficiency gains. To validate our theoretical findings, we conduct comprehensive synthetic experiments and apply our method to multiple real-world datasets, confirming the practical advantages of our approach.

Ask Lune

Know this paper in minutes, not an afternoon.

Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.

Questions to start from

Lune can draw on

the full paper · related work

Get the answer

Free to start. No credit card required.

Lineage

  • Identifiability of Label Noise Transition Matrix cites Provably End-to-end Label-noise Learning without Anchor Points
Builds onCited byLeft to right by year; the more cited, the higher.

Builds on6

  1. A Unified View of Label Shift EstimationSaurabh Garg, Yifan Wu, Sivaraman Balakrishnan, Zachary C. LiptonNeurIPS 2020 · 186 citations
  2. Provably End-to-end Label-noise Learning without Anchor PointsXuefeng Li, Tongliang Liu, Bo Han, Gang Niu et al.ICML 2021 · 161 citations
  3. Identifiability of Label Noise Transition MatrixYang Liu, Hao Cheng, Kun ZhangICML 2023 · 58 citations
  4. Label Correction of Crowdsourced Noisy Annotations with an Instance-Dependent Noise Transition ModelHui Guo, Boyu Wang, Grace YiNeurIPS 2023 · 21 citations
  5. Inferring the Long-Term Causal Effects of Long-Term Treatments from Short-Term ExperimentsAllen Tran, Aurélien Bibaut, Nathan KallusICML 2024 · 11 citations

Related papers

  1. Continuous Treatment Effects with Surrogate OutcomesZhenghao Zeng, David Arbour, Avi Feller, Raghavendra Addanki et al.ICML 2024 · 4 citations
  2. Complementary Benefits of Contrastive Learning and Self-Training Under Distribution ShiftSaurabh Garg, Amrith Setlur, Zachary C. Lipton, Sivaraman Balakrishnan et al.NeurIPS 2023 · 13 citations
  3. Out-Of-Domain Unlabeled Data Improves GeneralizationSeyed Amir Hossein Saberi, Amir Najafi, Alireza Heidari, Mohammad Hosein Movasaghinia et al.ICLR 2024 · 3 citations
  4. PUATE: Efficient ATE Estimation from Treated (Positive) and Unlabeled UnitsMasahiro Kato, Fumiaki Kozai, Ryo InokuchiNeurIPS 2025
  5. Doubly-Robust Self-TrainingBanghua Zhu, Mingyu Ding, Philip L. Jacobson, Ming Wu et al.NeurIPS 2023 · 22 citations
Dusk over the sea between two cliffs drawn in fine vertical lines