paper-with-me

홈 › Papers

Gradient Surgery for One-shot Unlearning on Generative Model

2023-07-10 · Seohui Bae, Seoyoon Kim, Hyemin Jung, Woohyung Lim

Recent regulation on right-to-be-forgotten emerges tons of interest in unlearning pre-trained machine learning models. While approximating a straightforward yet expensive approach of retrain-from-scratch, recent machine unlearning methods unlearn a sample by updating weights to remove its influence on the weight parameters. In this paper, we introduce a simple yet effective approach to remove a data influence on the deep generative model. Inspired by works in multi-task learning, we propose to manipulate gradients to regularize the interplay of influence among samples by projecting gradients onto the normal plane of the gradients to be retained. Our work is agnostic to statistics of the removal samples, outperforming existing baselines while providing theoretical analysis for the first time in unlearning a generative model.

📄 PDF Abstract BibTeX arXiv:2307.04550

Code (0)

등록된 구현이 없습니다.

Tasks

Machine UnlearningMulti-Task Learning

Similar Papers 제목 키워드 기반

Efficient Utility-Preserving Machine Unlearning with Implicit Gradient Surgery

2025-10-25 · Shiji Zhou, Tianbai Yu, Zhi Zhang, Heng Chang 외 arxiv

Machine unlearning (MU) aims to efficiently remove sensitive or harmful memory from a pre-trained model. The key challenge is to balance the potential tradeoff between unlearning efficacy and utility preservation, which …

UNO: Unlearning via Orthogonalization in Generative models

2025-06-05 · Pinak Mandal, Georg A. Gottwald

As generative models become increasingly powerful and pervasive, the ability to unlearn specific data, whether due to privacy concerns, legal requirements, or the correction of harmful content, has become increasingly im…

LLM Surgery: Efficient Knowledge Unlearning and Editing in Large Language Models

2024-09-19 · Akshaj Kumar Veldanda, Shi-Xiong Zhang, Anirban Das, Supriyo Chakraborty 외

Large language models (LLMs) have revolutionized various domains, yet their utility comes with significant challenges related to outdated or problematic knowledge embedded during pretraining. This paper addresses the cha…

Unlearning Concepts in Diffusion Model via Concept Domain Correction and Concept Preserving Gradient

2024-05-24 · Yongliang Wu, Shiji Zhou, Mingzhuo Yang, Lianzhe Wang 외

Text-to-image diffusion models have achieved remarkable success in generating photorealistic images. However, the inclusion of sensitive information during pre-training poses significant risks. Machine Unlearning (MU) of…

Image GenerationMachine Unlearning

Class Unlearning via Depth-Aware Removal of Forget-Specific Directions

2026-04-16 · Arman Hatami, Romina Aalishah, Ilya E. Monosov arxiv

Machine unlearning aims to remove targeted knowledge from a trained model without the cost of retraining from scratch. In class unlearning, however, reducing accuracy on forget classes does not necessarily imply true for…