mlsanityData to 14 Oct 2026

Anytime-valid, Bayes-assisted, Prediction-Powered Inference

arXiv 2505.18000 ↗First posted 23 May 2025stat.ML cs.LG math.STcited by 7

Not on the map. Why?

Abstract (arXiv)

Given a large pool of unlabelled data and a smaller amount of labels, prediction-powered inference (PPI) leverages machine learning predictions to increase the statistical efficiency of confidence interval procedures based solely on labelled data, while preserving fixed-time validity. In this paper, we extend the PPI framework to the sequential setting, where labelled and unlabelled datasets grow over time. Exploiting Ville's inequality and the method of mixtures, we propose prediction-powered confidence sequence procedures that are asymptotically valid uniformly over time and naturally accommodate prior knowledge on the quality of the predictions to further boost efficiency. We carefully illustrate the design choices behind our method and demonstrate its effectiveness in real and synthetic examples.

Key facts. Cited by 7 papers in the map (data through 14 Oct 2026). Builds on 19 of its 51 references that are in the map. First posted on arXiv 23 May 2025.

Builds on

19 of 51 references.

References · 19

  1. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  2. 2018Time-uniform, nonparametric, nonasymptotic confidence sequencescited by 229
  3. 2023Prediction-Powered Inferencecited by 220
  4. 2020Estimating means of bounded random variables by bettingcited by 169
  5. 2022Game-theoretic statistics and safe anytime-valid inferencecited by 155
  6. 2023PPI++: Efficient Prediction-Powered Inferencecited by 117
  7. 2018Mixture Martingales Revisited with Applications to Sequential Tests and Confidence Intervalscited by 76
  8. 2024Hypothesis testing with e-valuescited by 74
  9. 2024Active Statistical Inferencecited by 50
  10. 2016Semi-supervised Inference: General Theory and Estimation of Meanscited by 38
  11. 2021Time-uniform central limit theory and asymptotic confidence sequencescited by 29
  12. 2023A unified recipe for deriving (time-uniform) PAC-Bayes boundscited by 25
7 more
  1. 2022PAC-Bayes Generalisation Bounds for Heavy-Tailed Losses through Supermartingalescited by 24
  2. 2015Sequential Nonparametric Testing with the Law of the Iterated Logarithmcited by 20
  3. 2025Prediction-Powered E-Valuescited by 16
  4. 2022Near-Optimal Non-Parametric Sequential Tests and Confidence Sequences with Possibly Dependent Observationscited by 9
  5. 2025FAB-PPI: Frequentist, Assisted by Bayes, Prediction-Powered Inferencecited by 7
  6. 2024Anytime-Valid Inference for Double/Debiased Machine Learning of Causal Parameterscited by 5
  7. 2023The extended Ville's inequality for nonintegrable nonnegative supermartingalescited by 3

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 1; 2026: 6.

7 papers.

  1. 2026Who Drifted: the System or the Judge? Anytime-Valid Attribution in LLM Evaluation Pipelinescited by 8
  2. 2026Testing For Distribution Shifts with Conditional Conformal Test Martingalescited by 4
  3. 2026Prediction-Powered Risk Monitoring of Deployed Models for Detecting Harmful Distribution Shiftscited by 3
  4. 2025Differentially Private E-Valuescited by 1
  5. 2026Active Multiple-Prediction-Powered Inferencecited by 1
  6. 2026Semi-Supervised Hypothesis Testing by Betting on Predictionscited by 1
  7. 2026Asymptotically Log-Optimal Bayes-Assisted Confidence Sequences for Bounded Meanscited by 0