Saved in:
Bibliographic Details
Main Authors: Spartalis, Christoforos N., Semertzidis, Theodoros, Gavves, Efstratios, Daras, Petros
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2503.18314
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918131650265088
author Spartalis, Christoforos N.
Semertzidis, Theodoros
Gavves, Efstratios
Daras, Petros
author_facet Spartalis, Christoforos N.
Semertzidis, Theodoros
Gavves, Efstratios
Daras, Petros
contents We present LoTUS, a novel Machine Unlearning (MU) method that eliminates the influence of training samples from pre-trained models, avoiding retraining from scratch. LoTUS smooths the prediction probabilities of the model up to an information-theoretic bound, mitigating its over-confidence stemming from data memorization. We evaluate LoTUS on Transformer and ResNet18 models against eight baselines across five public datasets. Beyond established MU benchmarks, we evaluate unlearning on ImageNet1k, a large-scale dataset, where retraining is impractical, simulating real-world conditions. Moreover, we introduce the novel Retrain-Free Jensen-Shannon Divergence (RF-JSD) metric to enable evaluation under real-world conditions. The experimental results show that LoTUS outperforms state-of-the-art methods in terms of both efficiency and effectiveness. Code: https://github.com/cspartalis/LoTUS.
format Preprint
id arxiv_https___arxiv_org_abs_2503_18314
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle LoTUS: Large-Scale Machine Unlearning with a Taste of Uncertainty
Spartalis, Christoforos N.
Semertzidis, Theodoros
Gavves, Efstratios
Daras, Petros
Machine Learning
Artificial Intelligence
Computer Vision and Pattern Recognition
We present LoTUS, a novel Machine Unlearning (MU) method that eliminates the influence of training samples from pre-trained models, avoiding retraining from scratch. LoTUS smooths the prediction probabilities of the model up to an information-theoretic bound, mitigating its over-confidence stemming from data memorization. We evaluate LoTUS on Transformer and ResNet18 models against eight baselines across five public datasets. Beyond established MU benchmarks, we evaluate unlearning on ImageNet1k, a large-scale dataset, where retraining is impractical, simulating real-world conditions. Moreover, we introduce the novel Retrain-Free Jensen-Shannon Divergence (RF-JSD) metric to enable evaluation under real-world conditions. The experimental results show that LoTUS outperforms state-of-the-art methods in terms of both efficiency and effectiveness. Code: https://github.com/cspartalis/LoTUS.
title LoTUS: Large-Scale Machine Unlearning with a Taste of Uncertainty
topic Machine Learning
Artificial Intelligence
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2503.18314