Understanding Unfairness via Training Concept Influence

Yuanshun Yao, Yang Liu

Primary Area: societal considerations including fairness, safety, privacy

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.

Keywords: machine learning fairness; influence function; responsible AI

Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2024/AuthorGuide.

TL;DR: We propose a framework to help practitioners understand the observed unfairness and repair through training data.

Abstract: Knowing the causes of a model's unfairness helps practitioners better understand their data and algorithms. This is an important yet relatively unexplored task. We look into this problem through the lens of the training data - one of the major sources of unfairness. We ask the following questions: how a model's fairness performance would change if, in its training data, some samples (1) were collected from a different (e.g. demographic) group, (2) were labeled differently, or (3) some features were changed? In other words, we quantify the fairness influence of training samples by counterfactually changing samples based on predefined concepts, i.e. data attributes such as features, labels, or sensitive attributes. To calculate a training sample's influence on the model's unfairness w.r.t a concept, we first generate counterfactual samples based on the concept, i.e. the counterfactual versions of the sample if the concept were changed. We then calculate the resulting impact on the unfairness, via influence function, if the counterfactual samples were used in training. Our framework not only helps practitioners understand the observed unfairness and repair their training data, but also leads to many other applications, e.g. detecting mislabeling, fixing imbalanced representations, and detecting poisoning attacks.

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.

Submission Number: 3936