Saved in:
Bibliographic Details
Main Authors: Zhang, Yixuan, Zhou, Feng
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2403.00625
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929261445644288
author Zhang, Yixuan
Zhou, Feng
author_facet Zhang, Yixuan
Zhou, Feng
contents Fine-tuning pre-trained models is a widely employed technique in numerous real-world applications. However, fine-tuning these models on new tasks can lead to unfair outcomes. This is due to the absence of generalization guarantees for fairness properties, regardless of whether the original pre-trained model was developed with fairness considerations. To tackle this issue, we introduce an efficient and robust fine-tuning framework specifically designed to mitigate biases in new tasks. Our empirical analysis shows that the parameters in the pre-trained model that affect predictions for different demographic groups are different, so based on this observation, we employ a transfer learning strategy that neutralizes the importance of these influential weights, determined using Fisher information across demographic groups. Additionally, we integrate this weight importance neutralization strategy with a matrix factorization technique, which provides a low-rank approximation of the weight matrix using fewer parameters, reducing the computational demands. Experiments on multiple pre-trained models and new tasks demonstrate the effectiveness of our method.
format Preprint
id arxiv_https___arxiv_org_abs_2403_00625
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Bias Mitigation in Fine-tuning Pre-trained Models for Enhanced Fairness and Efficiency
Zhang, Yixuan
Zhou, Feng
Machine Learning
Computers and Society
Fine-tuning pre-trained models is a widely employed technique in numerous real-world applications. However, fine-tuning these models on new tasks can lead to unfair outcomes. This is due to the absence of generalization guarantees for fairness properties, regardless of whether the original pre-trained model was developed with fairness considerations. To tackle this issue, we introduce an efficient and robust fine-tuning framework specifically designed to mitigate biases in new tasks. Our empirical analysis shows that the parameters in the pre-trained model that affect predictions for different demographic groups are different, so based on this observation, we employ a transfer learning strategy that neutralizes the importance of these influential weights, determined using Fisher information across demographic groups. Additionally, we integrate this weight importance neutralization strategy with a matrix factorization technique, which provides a low-rank approximation of the weight matrix using fewer parameters, reducing the computational demands. Experiments on multiple pre-trained models and new tasks demonstrate the effectiveness of our method.
title Bias Mitigation in Fine-tuning Pre-trained Models for Enhanced Fairness and Efficiency
topic Machine Learning
Computers and Society
url https://arxiv.org/abs/2403.00625