Sample Efficient Learning of Predictors that Complement Humans
This work addresses the challenge of reducing human burden in decision-making by improving complementary learning, though it appears incremental as it builds on existing deferral frameworks.
The paper tackles the problem of learning predictors that complement human decision-makers in expert deferral settings, providing the first theoretical analysis of their benefit and designing active learning schemes that require minimal human prediction data.
One of the goals of learning algorithms is to complement and reduce the burden on human decision makers. The expert deferral setting wherein an algorithm can either predict on its own or defer the decision to a downstream expert helps accomplish this goal. A fundamental aspect of this setting is the need to learn complementary predictors that improve on the human's weaknesses rather than learning predictors optimized for average error. In this work, we provide the first theoretical analysis of the benefit of learning complementary predictors in expert deferral. To enable efficiently learning such predictors, we consider a family of consistent surrogate loss functions for expert deferral and analyze their theoretical properties. Finally, we design active learning schemes that require minimal amount of data of human expert predictions in order to learn accurate deferral systems.