文本到图像扩散模型的持续反学习：一种正则化视角
Continual Unlearning for Text-to-Image Diffusion Models: A Regularization Perspective
AI总结：
本文提出通过正则化方法解决文本到图像扩散模型中持续反学习的问题，通过减轻参数漂移和增强语义意识，提升反学习性能并推动生成式AI的安全发展。
摘要：
机器反学习——从预训练模型中移除指定概念的能力——已经迅速发展，尤其是在文本到图像扩散模型中。然而，现有方法通常假设反学习请求一次性到达，而在实践中它们往往依次到达。我们首次系统地研究了文本到图像扩散模型中的持续反学习，并表明流行的反学习方法存在快速的实用性崩溃：在仅几个请求后，模型会忘记保留的知识并生成退化图像。我们追溯这种失败到预训练权重的累积参数漂移，并认为正则化是解决这一问题的关键。为此，我们研究了一套附加的正则化器，这些正则化器（1）减轻漂移（2）保持与现有反学习方法的兼容性。除了通用正则化器之外，我们还显示语义意识对于保留接近反学习目标的概念至关重要，并提出了一种梯度投影方法，该方法约束参数漂移与其子空间正交。这大大提高了持续反学习的性能，并且与其他正则化器互补以获得进一步的收益。综合来看，我们的研究将持续反学习确立为文本到图像生成中的基本挑战，并提供了见解、基准和开放方向，以推动安全和负责任的生成式人工智能的发展。
英文摘要：
Machine unlearning--the ability to remove designated concepts from a pre-trained model--has advanced rapidly, particularly for text-to-image diffusion models. However, existing methods typically assume that unlearning requests arrive all at once, whereas in practice they often arrive sequentially. We present the first systematic study of continual unlearning in text-to-image diffusion models and show that popular unlearning methods suffer from rapid utility collapse: after only a few requests, models forget retained knowledge and generate degraded images. We trace this failure to cumulative parameter drift from the pre-training weights and argue that regularization is crucial to addressing it. To this end, we study a suite of add-on regularizers that (1) mitigate drift and (2) remain compatible with existing unlearning methods. Beyond generic regularizers, we show that semantic awareness is essential for preserving concepts close to the unlearning target, and propose a gradient-projection method that constrains parameter drift orthogonal to their subspace. This substantially improves continual unlearning performance and is complementary to other regularizers for further gains. Taken together, our study establishes continual unlearning as a fundamental challenge in text-to-image generation and provides insights, baselines, and open directions for advancing safe and accountable generative AI.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
