A Fairness Analysis on Differentially Private Aggregation of Teacher Ensembles

Cuong Tran, Ferdinando Fioretto

Keywords: Differential Privacy, Fairness, Semisupervised learning

TL;DR: This paper analyzes the causes of the disparate impacts arising in a popular teacher ensemble model used for differentially private learning tasks

Abstract:

Private Aggregation of Teacher Ensembles (PATE) is an important
private machine learning framework. It combines multiple
learning models used as teachers for a student model that
learns to predict an output chosen by noisy voting among the
teachers. The resulting model satisfies differential privacy and has
been shown effective in learning high-quality private models in
semi-supervised settings or when one wishes to protect the data
labels.

This paper asks whether this privacy-preserving framework introduces
or exacerbates unfairness and shows that PATE can introduce
accuracy disparity among individuals and groups of individuals.
The paper analyzes
which algorithmic and data properties are responsible for the
disproportionate impacts, why these aspects are affecting different
groups disproportionately, and proposes guidelines to mitigate these
effects.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

Submission Guidelines: Yes

Please Choose The Closest Area That Your Submission Falls Into: Social Aspects of Machine Learning (eg, AI safety, fairness, privacy, interpretability, human-AI interaction, ethics)