Identifying Bias in AI using Simulation

TL;DR

This paper introduces a simulation-based framework using Bayesian search to identify and diagnose demographic biases in machine learning classifiers, demonstrated on face detection APIs.

Contribution

It presents a novel approach leveraging high-fidelity simulations and Bayesian search to efficiently detect biases in ML models, improving bias diagnosis methods.

Findings

Effective identification of demographic biases in face detection APIs

Framework reduces time to diagnose biases compared to traditional methods

Demonstrates the utility of simulation in bias detection for ML models

Abstract

Machine learned models exhibit bias, often because the datasets used to train them are biased. This presents a serious problem for the deployment of such technology, as the resulting models might perform poorly on populations that are minorities within the training set and ultimately present higher risks to them. We propose to use high-fidelity computer simulations to interrogate and diagnose biases within ML classifiers. We present a framework that leverages Bayesian parameter search to efficiently characterize the high dimensional feature space and more quickly identify weakness in performance. We apply our approach to an example domain, face detection, and show that it can be used to help identify demographic biases in commercial face application programming interfaces (APIs).