A Comparison of Human and Machine Learning Errors in Face Recognition

TL;DR

This paper compares human and machine errors in face recognition through experiments, revealing differences in error patterns and suggesting strategies for improved human-machine collaboration.

Contribution

It provides a detailed comparison of human and machine face recognition errors, highlighting their differences and proposing ways to enhance accuracy through collaboration.

Findings

Machine learning and human errors differ significantly in face recognition.

Certain face recognition errors are more common in machines than humans, and vice versa.

Strategies for combining human and machine judgments can improve overall accuracy.

Abstract

Machine learning applications in high-stakes scenarios should always operate under human oversight. Developing an optimal combination of human and machine intelligence requires an understanding of their complementarities, particularly regarding the similarities and differences in the way they make mistakes. We perform extensive experiments in the area of face recognition and compare two automated face recognition systems against human annotators through a demographically balanced user study. Our research uncovers important ways in which machine learning errors and human errors differ from each other, and suggests potential strategies in which human-machine collaboration can improve accuracy in face recognition.