REVIEW 3 cited by
Evaluating Machine Unlearning via Epistemic Uncertainty
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
read the original abstract
There has been a growing interest in Machine Unlearning recently, primarily due to legal requirements such as the General Data Protection Regulation (GDPR) and the California Consumer Privacy Act. Thus, multiple approaches were presented to remove the influence of specific target data points from a trained model. However, when evaluating the success of unlearning, current approaches either use adversarial attacks or compare their results to the optimal solution, which usually incorporates retraining from scratch. We argue that both ways are insufficient in practice. In this work, we present an evaluation metric for Machine Unlearning algorithms based on epistemic uncertainty. This is the first definition of a general evaluation metric for Machine Unlearning to our best knowledge.
Forward citations
Cited by 3 Pith papers
-
Invisible Watermarks, Visible Gains: Steering Machine Unlearning with Bi-Level Watermarking Design
Water4MU tunes an invisible watermark on data so that machine unlearning algorithms can remove requested images more effectively, beating prior methods on 'challenging forgets'.
-
Recalling The Forgotten Class Memberships: Unlearned Models Can Be Noisy Labelers to Leak Privacy
An unlearned model can be used as a noisy teacher to reconstruct the class labels of data that machine unlearning was supposed to forget, without access to the original model.
-
Targeted Forgetting of Image Subgroups in CLIP Models
A three-stage forgetting, reminding, and restoring pipeline lets CLIP forget a targeted image subgroup without pre-training data while keeping zero-shot performance.
Discussion (0). Continue with ORCID to comment.