Empirically Validating Conformal Prediction on Modern Vision Architectures Under Distribution Shift and Long-tailed Data
Keywords: Conformal Prediction, Uncertainty estimation, Out-of-distribution
TL;DR: We test conformal prediction methods under distribution shift and in long-tailed datasets and show that coverage guarantees are violated and confidence sets grow.
Abstract: Conformal prediction has emerged as a rigorous means of providing deep learning models with reliable uncertainty estimates and safety guarantees. Yet, its performance is known to degrade under distribution shift and long-tailed class distributions, which are often present in real world applications. Here, we characterize the performance of several post-hoc and training-based conformal prediction methods under these settings, providing the first empirical evaluation on large-scale datasets and models. We show that across numerous conformal methods and neural network families, performance greatly degrades under distribution shifts violating safety guarantees. Similarly, we show that in long-tailed settings the guarantees are frequently violated on many classes. Understanding the limitations of these methods is necessary for deployment in real world and safety-critical applications.
Submission Number: 64
Loading