Proxy Non-Discrimination in Data-Driven Systems

Machine learnt systems inherit biases against protected classes, historically disparaged groups, from training data. Usually, these biases are not explicit, they rely on subtle correlations discovered by training algorithms, and are therefore difficult to detect. We formalize proxy discrimination in...

Full description

Saved in:  
Bibliographic Details
Authors: Datta, Anupam (Author) ; Sen, Shayak (Author) ; Mardziel, Piotr (Author) ; Ko, Gihyuk (Author) ; Fredrikson, Matt (Author)
Format: Electronic Book
Language:English
Published: 2017
In:Year: 2017
Online Access: Volltext (kostenfrei)
Check availability: HBZ Gateway
Keywords:
Research
Description
Summary:Machine learnt systems inherit biases against protected classes, historically disparaged groups, from training data. Usually, these biases are not explicit, they rely on subtle correlations discovered by training algorithms, and are therefore difficult to detect. We formalize proxy discrimination in data-driven systems, a class of properties indicative of bias, as the presence of protected class correlates that have causal influence on the system's output. We evaluate an implementation on a corpus of social datasets, demonstrating how to validate systems against these properties and to repair violations where they occur.Comment: arXiv admin note: substantial text overlap with arXiv:1705.0780