Evaluating Robustness to Dataset Shift via Parametric Robustness Sets

TL;DR

This paper introduces a method to identify and analyze small, plausible distribution shifts in data that can significantly impact model performance, using parametric robustness sets and second-order approximations.

Contribution

It proposes a novel approach to quantify model robustness to distribution shifts via parametric changes and efficient quadratic optimization techniques.

Findings

Method effectively detects sensitivity to distribution shifts in computer vision tasks.

Second-order approximation provides accurate worst-case loss estimates for small shifts.

Application reveals models' vulnerability to non-causal attribute shifts.

Abstract

We give a method for proactively identifying small, plausible shifts in distribution which lead to large differences in model performance. These shifts are defined via parametric changes in the causal mechanisms of observed variables, where constraints on parameters yield a "robustness set" of plausible distributions and a corresponding worst-case loss over the set. While the loss under an individual parametric shift can be estimated via reweighting techniques such as importance sampling, the resulting worst-case optimization problem is non-convex, and the estimate may suffer from large variance. For small shifts, however, we can construct a local second-order approximation to the loss under shift and cast the problem of finding a worst-case shift as a particular non-convex quadratic optimization problem, for which efficient algorithms are available. We demonstrate that this second-order approximation can be estimated directly for shifts in conditional exponential family models, and we bound the approximation error. We apply our approach to a computer vision task (classifying gender from images), revealing sensitivity to shifts in non-causal attributes.