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

arXiv: https://arxiv.org/abs/1907.12917 · HTML: https://mlsanity.com/paper/1907.12917/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 22 Jul 2019.

## Built on by (most cited first, 2 of 2)

- [StyleT2I: Toward Compositional and High-Fidelity Text-to-Image Synthesis](https://mlsanity.com/paper/2203.15799/) (2022)
- [FLUID: Training-Free Face De-identification via Latent Identity Substitution](https://mlsanity.com/paper/2511.17005/) (2025)

## Abstract (arXiv)

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.