Faster exact data deletion in federated fair clustering improves hospital and finance uses
Machine Unlearning for Federated Socially Fair k-Means Clustering
Machine LearningData Structures and Algorithms
Summary
Some computers work together on data stored in different places to group people into clusters, like patients or customers. This paper looks at how to remove data exactly when asked, such as under privacy laws, without starting the grouping from scratch. The authors focus on making sure groups are treated fairly and measure how well their method speeds up removal of data from whole hospitals or clients. They find big time savings when entire data sources are removed, but smaller gains for single data points. Their approach helps balance fairness and privacy in shared data clustering tasks.
What this means in practice
- •For hospital data teams: Enable efficient removal of an entire hospital’s data from shared patient clusters without full retraining, maintaining fair group representation.
- •For financial data analysts: Delete a client’s data from distributed customer segmentation models faster while preserving fairness across demographic groups in federated systems.
Authors
Aravind Reddy, Deeraj S K, Mohd Wasif Raza, Saurav Prakash
Abstract
Federated clustering can support applications such as patient stratification across hospitals and customer segmentation across distributed financial datasets. Responsible AI calls for fair representation of demographic groups and support for data deletion, motivated in part by rights such as the GDPR's right to erasure. These goals interact: deleting records can change the balance between groups and alter the clustering of the remaining population. We study exact machine unlearning for federated socially fair $k$-means, whose objective is to minimize the largest group-average squared distance to the nearest center. Our approach keeps separate local summaries for each group, so summaries at unaffected institutions remain valid as global fairness weights change. For two groups, we prove an initialization approximation guarantee independent of client group proportions under ideal sampling, and we show that same-seed replay reproduces retraining exactly. Reuse pays off most when a whole client withdraws: across three datasets, replay is $35.8\times$ faster than retraining for the initialization-only model and remains $2.4\times$ and $1.7\times$ faster with one and two fair-refinement rounds, which improve worst-group cost. For individual record deletions, gains are modest (about $1.06\times$), and certifying cached assignments can cost more than it saves. These results map where exact unlearning can profitably reuse computation. Code and reproducibility materials are available at https://github.com/arareddy/mufsc.