Targeted Unlearning with Single Layer Unlearning Gradient

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Cai, Zikui, Tan, Yaoteng, Asif, M. Salman
Formato: Preprint
Publicado: 2024
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866915312794861568
author Cai, Zikui
Tan, Yaoteng
Asif, M. Salman
author_facet Cai, Zikui
Tan, Yaoteng
Asif, M. Salman
contents Machine unlearning methods aim to remove sensitive or unwanted content from trained models, but typically demand extensive model updates at significant computational cost while potentially degrading model performance on both related and unrelated tasks. We propose Single Layer Unlearning Gradient (SLUG) as an efficient method to unlearn targeted information by updating a single critical layer using a one-time gradient computation. SLUG uses layer importance and gradient alignment metrics to identify the optimal layer for targeted information removal while preserving the model utility. We demonstrate the effectiveness of SLUG for CLIP, Stable Diffusion, and vision-language models (VLMs) in removing concrete (e.g., identities and objects) and abstract concepts (e.g., artistic styles). On the UnlearnCanvas benchmark, SLUG achieves comparable unlearning performance to existing methods while requiring significantly less computational resources. Our proposed approach offers a practical solution for targeted unlearning that is computationally efficient and precise. Our code is available at https://github.com/CSIPlab/SLUG.
format Preprint
id arxiv_https___arxiv_org_abs_2407_11867
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Targeted Unlearning with Single Layer Unlearning Gradient
Cai, Zikui
Tan, Yaoteng
Asif, M. Salman
Machine Learning
Machine unlearning methods aim to remove sensitive or unwanted content from trained models, but typically demand extensive model updates at significant computational cost while potentially degrading model performance on both related and unrelated tasks. We propose Single Layer Unlearning Gradient (SLUG) as an efficient method to unlearn targeted information by updating a single critical layer using a one-time gradient computation. SLUG uses layer importance and gradient alignment metrics to identify the optimal layer for targeted information removal while preserving the model utility. We demonstrate the effectiveness of SLUG for CLIP, Stable Diffusion, and vision-language models (VLMs) in removing concrete (e.g., identities and objects) and abstract concepts (e.g., artistic styles). On the UnlearnCanvas benchmark, SLUG achieves comparable unlearning performance to existing methods while requiring significantly less computational resources. Our proposed approach offers a practical solution for targeted unlearning that is computationally efficient and precise. Our code is available at https://github.com/CSIPlab/SLUG.
title Targeted Unlearning with Single Layer Unlearning Gradient
topic Machine Learning
url https://arxiv.org/abs/2407.11867