Covering up bias in CelebA-like datasets with Markov blankets: A post-hoc cure for attribute prior avoidance

Attribute prior avoidance entails subconscious or willful non-modeling of (meta)attributes that datasets are oft born with, such as the 40 semantic facial attributes associated with the CelebA and CelebA-HQ datasets. The consequences of this infirmity, we discover, are especially stark in state-of-t...

Full description

Saved in:
Bibliographic Details
Published in:arXiv.org 2019-07
Main Authors: Prabhu, Vinay Uday, Dian Ang Yap, Wang, Alexander, Whaley, John
Format: Article
Language:English
Subjects:
Avoidance
Bias
Datasets
Ising model
Markov processes
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Attribute prior avoidance entails subconscious or willful non-modeling of (meta)attributes that datasets are oft born with, such as the 40 semantic facial attributes associated with the CelebA and CelebA-HQ datasets. The consequences of this infirmity, we discover, are especially stark in state-of-the-art deep generative models learned on these datasets that just model the pixel-space measurements, resulting in an inter-attribute bias-laden latent space. This viscerally manifests itself when we perform face manipulation experiments based on latent vector interpolations. In this paper, we address this and propose a post-hoc solution that utilizes an Ising attribute prior learned in the attribute space and showcase its efficacy via qualitative experiments.
ISSN:2331-8422