mlsanityData to 14 Oct 2026

Learning-To-Measure: In-Context Active Feature Acquisition

arXiv 2510.12624 ↗First posted 14 Oct 2025cs.LG cs.AIcited by 5

Not on the map. Why?

Abstract (arXiv)

Active feature acquisition (AFA) is a sequential decision-making problem where the goal is to improve model performance for test instances by adaptively selecting which features to acquire. In practice, AFA methods often learn from retrospective data with systematic missingness in the features and limited task-specific labels. Most prior work addresses acquisition for a single predetermined task, limiting scalability. To address this limitation, we formalize the meta-AFA problem, where the goal is to learn acquisition policies across various tasks. We introduce Learning-to-Measure (L2M), which consists of i) reliable uncertainty quantification over unseen tasks, and ii) an uncertainty-guided greedy feature acquisition agent that maximizes conditional mutual information. We demonstrate a sequence-modeling or autoregressive pre-training approach that underpins reliable uncertainty quantification for tasks with arbitrary missingness. L2M operates directly on datasets with retrospective missingness and performs the meta-AFA task in-context, eliminating per-task retraining. Across synthetic and real-world tabular benchmarks, L2M matches or surpasses task-specific baselines, particularly under scarce labels and high missingness.

Key facts. Cited by 5 papers in the map (data through 14 Oct 2026). Builds on 39 of its 62 references that are in the map. First posted on arXiv 14 Oct 2025.

Builds on

39 of 62 references.

References · 39

  1. 2017Attention Is All You Needcited by 60,361
  2. 2017Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networkscited by 4,897
  3. 2021On the Opportunities and Risks of Foundation Modelscited by 4,033
  4. 2016The Concrete Distribution: A Continuous Relaxation of Discrete Random Variablescited by 1,148
  5. 2015Dueling Network Architectures for Deep Reinforcement Learningcited by 785
  6. 2022TabPFN: A Transformer That Solves Small Tabular Classification Problems in a Secondcited by 674
  7. 2018Conditional Neural Processescited by 426
  8. 2019Efficient Off-Policy Meta-Reinforcement Learning via Probabilistic Context Variablescited by 389
  9. 2021Transformers Can Do Bayesian Inferencecited by 360
  10. 2019Monte Carlo Gradient Estimation in Machine Learningcited by 283
  11. 2023Supervised Pretraining Can Learn In-Context Reinforcement Learningcited by 139
  12. 2022Transformer Neural Processes: Uncertainty-Aware Meta Learning Via Sequence Modelingcited by 127
27 more
  1. 2018EDDI: Efficient Dynamic Discovery of High-Value Information with Partial VAEcited by 87
  2. 2023Transformers as Decision Makers: Provable In-Context Reinforcement Learning via Supervised Pretrainingcited by 83
  3. 2024Large Scale Transfer Learning for Tabular Data via Language Modelingcited by 76
  4. 2023Statistical Foundations of Prior-Data Fitted Networkscited by 65
  5. 2020Identifying Causal-Effect Inference Failure with Uncertainty-Aware Modelscited by 57
  6. 2023Learning to Maximize Mutual Information for Dynamic Feature Selectioncited by 56
  7. 2020Active Feature Acquisition with Generative Surrogate Modelscited by 54
  8. 2025Real-TabPFN: Improving Tabular Foundation Models via Continued Pre-training With Real-World Datacited by 41
  9. 2017Classification with Costly Features using Deep Reinforcement Learningcited by 34
  10. 2021Identifiable Generative Models for Missing Not at Random Data Imputationcited by 30
  11. 2022Interpretable by Design: Learning Predictors by Composing Interpretable Queriescited by 28
  12. 2020Full Law Identification In Graphical Models Of Missing Data: Completeness Resultscited by 26
  13. 2023Variational Information Pursuit for Interpretable Predictionscited by 25
  14. 2023Estimating Conditional Mutual Information for Dynamic Feature Selectioncited by 21
  15. 2019Opportunistic Learning: Budgeted Cost-Sensitive Learning from Data Streamscited by 20
  16. 2020Empirical Frequentist Coverage of Deep Learning Uncertainty Quantification Procedurescited by 19
  17. 2022Missing Data Imputation and Acquisition with Deep Hierarchical Models and Hamiltonian Monte Carlocited by 19
  18. 2024Meta-DT: Offline Meta-RL as Conditional Sequence Modeling with World Model Disentanglementcited by 17
  19. 2023Evaluation of Active Feature Acquisition Methods for Time-varying Feature Settingscited by 9
  20. 2024Active Exploration via Autoregressive Generation of Missing Datacited by 9
  21. 2025Stochastic Encodings for Active Feature Acquisitioncited by 9
  22. 2024Exchangeable Sequence Models Quantify Uncertainty Over Latent Conceptscited by 7
  23. 2022Imputation Strategies Under Clinical Presence: Impact on Algorithmic Fairnesscited by 6
  24. 2023Evaluation of Active Feature Acquisition Methods for Static Feature Settingscited by 4
  25. 2024Distribution Guided Active Feature Acquisitioncited by 3
  26. 2024Access to care improves EHR reliability and clinical risk prediction model performancecited by 3
  27. 2025Architectural and Inferential Inductive Biases For Exchangeable Sequence Modelingcited by 2

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 5.

5 papers.

  1. 2026Zero-Shot Active Feature Acquisition via LLM-Elicitationcited by 0
  2. 2026Wide Learning: Learning to Reach Evidencecited by 0
  3. 2026When Does Test-Time Physical Diagnosis Pay? A Frozen Policy Buys Evidence It Never Readscited by 0
  4. 2026Active Feature Acquisition With Incomplete Training Datacited by 0
  5. 2026Measure Less, Know More: Self-Supervised Test-Time Feature Acquisitioncited by 0