Not All Mislabels Are Equal: Efficient Task-Aware Auditing with Confidence Sequences
In submission; preliminary version in NeurIPS 2026 Workshop “E-Values: From Statistics to ML”, 2026
Fangyuan Lin, Noah Krever, Jakub Cerný, Lily Xu
Research
My research connects probability theory with statistical inference: concentration under dependence, selection bias in random walks, and adaptive sequential methods for auditing tasks.
You can also find my articles on my Google Scholar profile.
Publications and Preprints
The Winner’s Curse as an e-Process Compensator with Applications to Safe Anytime-Valid Testing for Dose-Ranging Trials
arXiv preprint, 2026
Bounded Difference Concentration for Infinitely Exchangeable Sequences with Applications to AI Benchmark Uncertainty
arXiv preprint, 2026
A Selection Premium Decomposition for the Expected Maximum of Random Walks
arXiv preprint, 2026
Try telescoping
An analysis of silk density in spider webs
Royal Society Open Science, 2025
“If you know your neighbors, do you know yourself?
Revisiting the Unicity Distance through a Channel Transmission Perspective
To appear in the PUMP Journal of Undergraduate Research (status in October 2026 CV), 2024
Fangyuan Lin