Understanding Fixed Predictions via Confined Regions

TL;DR

This paper introduces a new method to identify fixed predictions in machine learning models by finding confined regions in feature space, enabling better certification of recourse and out-of-sample prediction analysis.

Contribution

It proposes a novel paradigm for detecting fixed predictions through confined regions, overcoming limitations of pointwise methods and providing interpretable insights.

Findings

Existing pointwise methods fail to predict fixed predictions in new data.

The proposed method effectively identifies fixed prediction regions in diverse applications.

Confined region approach offers interpretable descriptions of fixed predictions.

Abstract

Machine learning models can assign fixed predictions that preclude individuals from changing their outcome. Existing approaches to audit fixed predictions do so on a pointwise basis, which requires access to an existing dataset of individuals and may fail to anticipate fixed predictions in out-of-sample data. This work presents a new paradigm to identify fixed predictions by finding confined regions of the feature space in which all individuals receive fixed predictions. This paradigm enables the certification of recourse for out-of-sample data, works in settings without representative datasets, and provides interpretable descriptions of individuals with fixed predictions. We develop a fast method to discover confined regions for linear classifiers using mixed-integer quadratically constrained programming. We conduct a comprehensive empirical study of confined regions across diverse applications. Our results highlight that existing pointwise verification methods fail to anticipate future individuals with fixed predictions, while our method both identifies them and provides an interpretable description.