Obliviate: 从自回归图像生成模型中擦除概念
Obliviate: Erasing Concepts from Autoregressive Image Generation Models
AI总结:
提出Obliviate方法,通过KL散度监督视觉token分布、轨迹级更新和对齐视觉前缀,有效擦除自回归文生图模型中的有害概念,在保持模型效用同时大幅降低不当内容生成。
摘要:
生成式AI模型的广泛采用加剧了对其滥用的担忧,包括创建不安全或令人不安的图像。为了缓解此类问题,已有多种概念擦除方法被提出以从多模态生成模型中移除有害内容。然而,尽管自回归图像生成模型在统一多模态架构的近期趋势中日益重要,其概念擦除仍基本未被探索。在本工作中,我们通过引入Obliviate填补了这一空白,这是一种基于引导的自回归图像生成概念擦除方法。我们的方法基于三个关键设计选择:基于KL散度的视觉token分布监督、完整自回归展开上的轨迹级更新,以及用于稳定目标构建的对齐视觉前缀。我们在三种最先进的自回归文生图模型——Liquid、Emu3-Gen和Janus-Pro上评估了Obliviate,涵盖了显式内容、图形暴力和品牌图像的擦除。Obliviate始终优于现有替代方案,在防御性RAB基准上将裸体内容从91.58降至3.15,同时保持了整体模型效用。
英文摘要:
The widespread adoption of generative AI models has intensified concerns about misuse, including the creation of unsafe or disturbing imagery. To mitigate such issues, several concept erasure approaches have been proposed to remove harmful content from multimodal generative models. Yet concept erasure for autoregressive image generation remains largely unexplored, despite the growing relevance of these models in recent trends toward unified multimodal architectures. In this work, we fill this gap by introducing Obliviate, a guidance-based concept erasure method for autoregressive image generation. Our method builds on three key design choices: KL-based supervision over visual token distributions, trajectory-level updates over full autoregressive rollouts, and aligned visual prefixes for stable target construction. We evaluate Obliviate on three state-of-the-art autoregressive text-to-image models, Liquid, Emu3-Gen, and Janus-Pro, covering the erasure of explicit content, graphic violence, and branded imagery. Obliviate consistently outperforms current alternatives, reducing nudity on the defensive RAB benchmark from 91.58 to 3.15 while preserving overall model utility. Our code is available at https://github.com/multimodal-ai-lab/Obliviate.
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。