Derandomized Novelty Detection with FDR Control via Conformal E-values
This work addresses the issue of interpretability in novelty detection for researchers and practitioners by providing a more stable method, though it is incremental as it builds on existing conformal inference frameworks.
The paper tackles the problem of randomness in conformal inference for novelty detection by introducing a method using conformal e-values to aggregate evidence and control the false discovery rate, showing reduced randomness with minimal power loss compared to standard techniques.
Conformal inference provides a general distribution-free method to rigorously calibrate the output of any machine learning algorithm for novelty detection. While this approach has many strengths, it has the limitation of being randomized, in the sense that it may lead to different results when analyzing twice the same data, and this can hinder the interpretation of any findings. We propose to make conformal inferences more stable by leveraging suitable conformal e-values instead of p-values to quantify statistical significance. This solution allows the evidence gathered from multiple analyses of the same data to be aggregated effectively while provably controlling the false discovery rate. Further, we show that the proposed method can reduce randomness without much loss of power compared to standard conformal inference, partly thanks to an innovative way of weighting conformal e-values based on additional side information carefully extracted from the same data. Simulations with synthetic and real data confirm this solution can be effective at eliminating random noise in the inferences obtained with state-of-the-art alternative techniques, sometimes also leading to higher power.