为二进制数据设计的生成模型，速度快效果好。
Binary Diffusion Probabilistic Model
- 用XOR加噪和二元交叉熵训练，适配二进制表示。
- 少步采样下在超分、修复等任务上超越现有方法。
- 适合低参数量、快速生成场景，如边缘设备部署。
我们提出二进制扩散概率模型（BDPM），一种专为二进制数据表示设计的生成框架。传统去噪扩散概率模型（DDPM）假设输入为连续值，使用均方误差目标和高斯噪声，这些假设不适用于离散二进制表示。BDPM通过多比特平面与可学习二进制嵌入将图像编码为二进制表示，采用基于XOR的噪声扰动，并通过优化二元交叉熵损失进行训练。该二进制表示提供了细粒度噪声控制，加速收敛并降低推理成本。在图像到图像翻译任务（如超分辨率、修复和盲恢复）中，基于小型去噪器和多比特平面表示的BDPM，在FFHQ、CelebA和CelebA-HQ数据集上仅用少量采样步骤即优于当前最优方法。在ImageNet-1k上的类别条件生成任务中，基于可学习二进制嵌入的BDPM在低参数量和少采样步数条件下也达到了具有竞争力的生成效果。
原文摘要 · Abstract (English)
We propose the Binary Diffusion Probabilistic Model (BDPM), a generative framework specifically designed for data representations in binary form. Conventional denoising diffusion probabilistic models (DDPMs) assume continuous inputs, use mean squared error objectives and Gaussian perturbations, i.e., assumptions that are not suited to discrete and binary representations. BDPM instead encodes images into binary representations using multi bit-plane and learnable binary embeddings, perturbs them via XOR-based noise, and trains a model by optimizing a binary cross-entropy loss. These binary representations offer fine-grained noise control, accelerate convergence, and reduce inference cost. On image-to-image translation tasks, such as super-resolution, inpainting, and blind restoration, BDPM based on a small denoiser and multi bit-plane representation outperforms state-of-the-art methods on FFHQ, CelebA, and CelebA-HQ using a few sampling steps. In class-conditional generation on ImageNet-1k, BDPM based on learnable binary embeddings achieves competitive results among models with both low parameter counts and a few sampling steps.
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。