A Causal Framework for Evaluating Deferring Systems

TL;DR

This paper introduces a causal inference framework to evaluate the impact of deferring strategies in machine learning systems, addressing a gap in understanding how deferring affects accuracy.

Contribution

It links causal inference methods with deferring systems, enabling causal impact assessment in scenarios with varying data availability.

Findings

Effective causal estimation methods for deferring systems.

Application to synthetic and real datasets demonstrates practical utility.

Distinction between scenarios with and without human predictions.

Abstract

Deferring systems extend supervised Machine Learning (ML) models with the possibility to defer predictions to human experts. However, evaluating the impact of a deferring strategy on system accuracy is still an overlooked area. This paper fills this gap by evaluating deferring systems through a causal lens. We link the potential outcomes framework for causal inference with deferring systems, which allows to identify the causal impact of the deferring strategy on predictive accuracy. We distinguish two scenarios. In the first one, we have access to both the human and ML model predictions for the deferred instances. Here, we can identify the individual causal effects for deferred instances and the aggregates of them. In the second one, only human predictions are available for the deferred instances. Here, we can resort to regression discontinuity designs to estimate a local causal effect. We evaluate our approach on synthetic and real datasets for seven deferring systems from the literature.