When unlearning is free: leveraging low influence points to reduce computational costs

TL;DR

This paper introduces a method to identify and remove low-impact data points to efficiently unlearn information from models, significantly reducing computational costs in privacy-sensitive machine learning applications.

Contribution

The authors propose a novel unlearning framework that leverages influence functions to identify negligible-impact points, enabling dataset reduction and computational savings.

Findings

Up to 50% reduction in unlearning computational costs

Effective identification of low-impact data points across tasks

Significant efficiency gains demonstrated on real-world data

Abstract

As concerns around data privacy in machine learning grow, the ability to unlearn, or remove, specific data points from trained models becomes increasingly important. While state of the art unlearning methods have emerged in response, they typically treat all points in the forget set equally. In this work, we challenge this approach by asking whether points that have a negligible impact on the model's learning need to be removed. Through a comparative analysis of influence functions across language and vision tasks, we identify subsets of training data with negligible impact on model outputs. Leveraging this insight, we propose an efficient unlearning framework that reduces the size of datasets before unlearning leading to significant computational savings (up to approximately 50 percent) on real world empirical examples.