Decision-Focused Evaluation of Worst-Case Distribution Shift

TL;DR

This paper introduces a new framework for identifying worst-case distribution shifts in predictive resource allocation, focusing on population-level decision impacts rather than individual accuracy, and demonstrates its effectiveness on real data.

Contribution

The paper proposes a hierarchical model-based approach to detect worst-case shifts in decision-making contexts, reformulating the problem as a submodular optimization for efficient solutions.

Findings

Worst-case shifts vary significantly across different metrics.

The framework effectively captures complex distribution shifts in real data.

Identifies shifts that traditional accuracy-based methods may overlook.

Abstract

Distribution shift is a key challenge for predictive models in practice, creating the need to identify potentially harmful shifts in advance of deployment. Existing work typically defines these worst-case shifts as ones that most degrade the individual-level accuracy of the model. However, when models are used to make a downstream population-level decision like the allocation of a scarce resource, individual-level accuracy may be a poor proxy for performance on the task at hand. We introduce a novel framework that employs a hierarchical model structure to identify worst-case distribution shifts in predictive resource allocation settings by capturing shifts both within and across instances of the decision problem. This task is more difficult than in standard distribution shift settings due to combinatorial interactions, where decisions depend on the joint presence of individuals in the allocation task. We show that the problem can be reformulated as a submodular optimization problem, enabling efficient approximations of worst-case loss. Applying our framework to real data, we find empirical evidence that worst-case shifts identified by one metric often significantly diverge from worst-case distributions identified by other metrics.