Forget Vectors at Play: Universal Input Perturbations Driving Machine Unlearning in Image Classification
Keywords: Machine Unlearning, Image Classification, Universal Input Perturbations
Abstract: Machine unlearning (MU), which seeks to erase the influence of specific unwanted data from already-trained models, is becoming increasingly vital in model editing, particularly to comply with evolving data regulations like the "right to be forgotten''. Conventional approaches are predominantly model-based, typically requiring retraining or fine-tuning the model's weights to meet unlearning requirements. In this work, we approach the MU problem from a novel input perturbation-based perspective, where the model weights remain intact throughout the unlearning process. We demonstrate the existence of a proactive input-based unlearning strategy, referred to forget vector, which can be generated as an input-agnostic data perturbation and remains as effective as model-based approximate unlearning approaches. We also show that multiple given forget vectors (e.g., each targeting the unlearning of a specific data class) can be combined through simple arithmetic operations (e.g., linear combinations) to generate new forget vectors for unseen unlearning tasks (e.g., targeting the unlearning of an arbitrary subset across all classes). An additional advantage of our proposed forget vector approach is its parameter efficiency, as it eliminates the need for updating model weights. We conduct extensive experiments to validate the effectiveness of forget vector and its arithmetic for MU in image classification against a series of model-based unlearning baselines.
Primary Area: alignment, fairness, safety, privacy, and societal considerations
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.
Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.
Reciprocal Reviewing: I understand the reciprocal reviewing requirement as described on https://iclr.cc/Conferences/2025/CallForPapers. If none of the authors are registered as a reviewer, it may result in a desk rejection at the discretion of the program chairs. To request an exception, please complete this form at https://forms.gle/Huojr6VjkFxiQsUp6.
Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.
No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.
Submission Number: 5198
Loading