CVAILGMar 30, 2023

Forget-Me-Not: Learning to Forget in Text-to-Image Diffusion Models

Georgia Tech
arXiv:2303.17591v1324 citationsh-index: 81Has Code
Originality Incremental advance
AI Analysis

It addresses a critical industry issue for developers and users of generative AI models by enabling safe and efficient concept removal, though it is incremental as it builds on existing unlearning methods.

The paper tackles the problem of unlearning specific concepts like IDs, objects, or styles from text-to-image diffusion models to address privacy, copyright, and safety concerns, achieving removal in as little as 30 seconds without impairing other content generation.

The unlearning problem of deep learning models, once primarily an academic concern, has become a prevalent issue in the industry. The significant advances in text-to-image generation techniques have prompted global discussions on privacy, copyright, and safety, as numerous unauthorized personal IDs, content, artistic creations, and potentially harmful materials have been learned by these models and later utilized to generate and distribute uncontrolled content. To address this challenge, we propose \textbf{Forget-Me-Not}, an efficient and low-cost solution designed to safely remove specified IDs, objects, or styles from a well-configured text-to-image model in as little as 30 seconds, without impairing its ability to generate other content. Alongside our method, we introduce the \textbf{Memorization Score (M-Score)} and \textbf{ConceptBench} to measure the models' capacity to generate general concepts, grouped into three primary categories: ID, object, and style. Using M-Score and ConceptBench, we demonstrate that Forget-Me-Not can effectively eliminate targeted concepts while maintaining the model's performance on other concepts. Furthermore, Forget-Me-Not offers two practical extensions: a) removal of potentially harmful or NSFW content, and b) enhancement of model accuracy, inclusion and diversity through \textbf{concept correction and disentanglement}. It can also be adapted as a lightweight model patch for Stable Diffusion, allowing for concept manipulation and convenient distribution. To encourage future research in this critical area and promote the development of safe and inclusive generative models, we will open-source our code and ConceptBench at \href{https://github.com/SHI-Labs/Forget-Me-Not}{https://github.com/SHI-Labs/Forget-Me-Not}.

Code Implementations1 repo
Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes