提出新训练方法，让扩散模型生成更原创图像且不依赖记忆数据。
Does Generation Require Memorization? Creative Diffusion Models using Ambient Diffusion
- 在高噪声尺度下训练扩散模型，避免低噪声时的记忆倾向。
- 实验显示生成质量不变，但记忆训练数据的程度显著降低。
- 适合追求生成多样性与隐私保护的图像生成研究者。
现有最先进的扩散模型在小数据集上容易记忆训练样本，现有缓解方法常以牺牲图像质量为代价。本文首次从理论上证明：记忆行为仅在低噪声尺度（用于生成高频细节）的去噪任务中必要。基于此，我们提出一种简单而合理的方法——在大噪声尺度下使用含噪数据训练扩散模型。实验证明，该方法在文本条件与无条件生成任务中，均能显著降低模型对训练数据的记忆程度，同时保持图像生成质量，适用于多种数据可用性场景。
原文摘要 · Abstract (English)
There is strong empirical evidence that the state-of-the-art diffusion modeling paradigm leads to models that memorize the training set, especially when the training set is small. Prior methods to mitigate the memorization problem often lead to a decrease in image quality. Is it possible to obtain strong and creative generative models, i.e., models that achieve high generation quality and low memorization? Despite the current pessimistic landscape of results, we make significant progress in pushing the trade-off between fidelity and memorization. We first provide theoretical evidence that memorization in diffusion models is only necessary for denoising problems at low noise scales (usually used in generating high-frequency details). Using this theoretical insight, we propose a simple, principled method to train the diffusion models using noisy data at large noise scales. We show that our method significantly reduces memorization without decreasing the image quality, for both text-conditional and unconditional models and for a variety of data availability settings.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。