Joint Transfer of Model Knowledge and Fairness Over Domains Using Wasserstein Distance

Taeho Yoon,Woojin Lee,Jaewook Lee

doi:10.1109/access.2020.3005987

Taeho Yoon, Woojin Lee + Show 1 more

Open Access

https://doi.org/10.1109/access.2020.3005987

Copy DOI

Abstract

Owing to the increasing use of machine learning in our daily lives, the problem of fairness has recently become an important topic in machine learning societies. Recent studies regarding fairness in machine learning have been conducted to attempt to ensure statistical independence between individual model predictions and designated sensitive attributes. However, in reality, cases exist in which the sensitive variables of data used for learning models differ from the data upon which the model is applied. In this paper, we investigate a methodology for developing a fair classification model for data with limited or no labels, by transferring knowledge from another data domain where information is fully available. This is done by controlling the Wasserstein distances between relevant distributions. Subsequently, we obtain a fair model that could be successfully applied to two datasets with different sensitive attributes. We present theoretical results validating that our approach provably transfers both classification performance and fairness over domains. Experimental results show that our method does indeed promote fairness for the target domain, while retaining reasonable classification accuracy, and that it often outperforms comparative models in terms of joint fairness.

Highlights

Machine learning is widely used in a variety of decision-making scenarios such as health care, criminal risk assessment, and financial lending
We introduce the notion of Strong Pairwise Demographic Disparity (SPDD), which has been originally proposed by [7], and Strong Pairwise Disparity of Opportunity (SPDOp), which is a corresponding concept for the criterion Equal opportunities (EOp)
We applied the concept of domain adaptation for fair machine learning to address the problem

Summary

Introduction

Machine learning is widely used in a variety of decision-making scenarios such as health care, criminal risk assessment, and financial lending. As machine learning is permeating our everyday lives, its fairness is becoming a real issue, and researchers are investigating the problem. Machine learning models are trained to predict outcomes for new samples using a given set of labeled examples. Traditional machine learning models for classification were designed to maximize the accuracy of their predictions; accurate predictions may still be unfair. This led to the growth of the literature on fairness in machine learning. These works consider one of the following two types: The associate editor coordinating the review of this manuscript and approving it for publication was Long Cheng

Objectives

Methods

Results

Conclusion