5 papers
Mitigating the Winner's Curse While Controlling Multiplicity: e-Process Methods for Anytime-Valid Inference in Dose-Ranging Trials
Victor K. de la Pena, Fangyuan Lin, Demissie Alemayehu +1
Phase II dose-ranging trials often report the largest observed dose-control effect while inspecting accumulating data repeatedly. This creates two coupled distortions: selection op…
Bounded Difference Concentration for Infinitely Exchangeable Sequences with Applications to AI Benchmark Uncertainty
Fangyuan Lin, Spencer Frei, Victor H. de la Pena
We consider the concentration properties of functions of infinitely exchangeable random variables. By conditioning on the de Finetti directing measure, we show that the deviation o…
A Selection Premium Decomposition for the Expected Maximum of Random Walks
Victor H. de la Pena, Fangyuan Lin, Victor K. de la Pena
When models are evaluated on the same validation set of size , the selected winner's apparent performance is biased upward. Suppose models are evaluated on a shared sequ…
Revisiting the Unicity Distance through a Channel Transmission Perspective
Fangyuan Lin
This paper revisits the classical notion of unicity distance from an enlightening perspective grounded in information theory, specifically by framing the encryption process as a no…
Applications of the Theory of Aggregated Markov Processes in Stochastic Learning Theory
Fangyuan Lin
A stochastic process that arises by composing a function with a Markov process is called an aggregated Markov process (AMP). The purpose of composing a Markov process with a functi…