Online Selective Conformal Prediction with Asymmetric Rules: A Permutation Test Approach
Mingyi Zheng and Ying Jin · 2026
Read Online Selective Conformal Prediction with Asymmetric Rules: A Permutation Test Approach by Mingyi Zheng and Ying Jin (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-10018.txt.
Description
Prediction sets are easy to trust when the test case was chosen blindly, but selection changes the odds. A 2026 paper by Mingyi Zheng and Ying Jin studies selective conformal prediction online, where data arrives in sequence and later choices depend on earlier ones. Existing offline methods assume selection treats labeled data symmetrically, an assumption online selection breaks.
The authors propose PEMI, a permutation-based framework that works for arbitrary asymmetric selection rules. It finds every reordering of the data that preserves the selection event, then calibrates the prediction set on that reference set. Under standard exchangeability, the sets achieve exact finite-sample coverage conditional on selection. Tests on a drug discovery data set and in simulation support the method.
Catalog record
| CREATOR | Mingyi Zheng and Ying Jin |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 137 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-10018