发表机构
Wrocław University of Science and Technology; IDEAS Research Institute; Jagiellonian University; AKCES NCBR; Tooploox（弗罗茨瓦夫理工大学; IDEAS研究所; 雅盖隆大学; AKCES NCBR; Tooploox）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
提出CLASP，用单一固定超网络持续个性化扩散模型，无需重放，参数占用不随概念数增长，并集成空间控制，实验证明保留旧概念且空间定位可靠。
AI 中文摘要
文本到图像扩散模型的持续个性化需要按顺序获取新概念，同时保留先前学到的概念。然而，现有方法要么遭受灾难性遗忘，要么依赖存储额外的概念特定参数和空间组件，导致其参数占用随概念流增长。这限制了它们扩展到长序列个性化任务的能力。我们提出一种无需重放的方法，使用单一固定大小的超网络持续个性化冻结的扩散模型。超网络不随新概念的获取而扩展模型，而是动态生成个性化所需的概念特定适配，同时保留先前学到的概念。我们的框架进一步将空间控制集成到个性化过程中，允许用户指定个性化概念应出现的位置，而无需引入额外的逐概念组件。这一公式使得持续个性化的参数占用与学习概念数量无关（除紧凑的概念表示外）。实验表明，该方法能强保留先前学到的概念，并提供可靠的空间定位，在匹配或改进现有方法的同时，有效扩展到长序列的个性化任务。
英文摘要
Continual personalization of text-to-image diffusion models requires sequentially acquiring new concepts while retaining previously learned ones. However, existing methods either suffer from catastrophic forgetting or rely on storing additional concept-specific parameters and spatial components, causing their parameter footprint to grow with the concept stream. This limits their ability to scale to long sequences of personalization tasks. We propose a rehearsal-free approach that uses a single fixed-size hypernetwork to continually personalize a frozen diffusion model. Instead of expanding the model as new concepts are acquired, the hypernetwork dynamically produces the concept-specific adaptations required for personalization while preserving previously learned concepts. Our framework further integrates spatial control into the personalization process, allowing users to specify where a personalized concept should appear without introducing additional per-concept components. This formulation enables continual personalization with a parameter footprint that remains independent of the number of learned concepts, aside from compact concept representations. Experiments demonstrate strong retention of previously learned concepts and reliable spatial grounding, matching or improving upon existing methods while scaling effectively to long streams of personalization tasks.
Comments31 pages. Code: https://github.com/genwro-ai/clasp, project page: https://genwro-ai.github.io/clasp
