DMGD：扩散模型中基于语义-分布匹配的无训练数据集蒸馏
DMGD: Train-Free Dataset Distillation with Semantic-Distribution Matching in Diffusion Models
用条件似然优化替代辅助分类器实现语义匹配，并以最优传输做分布匹配，为扩散模型中的无训练引导提供可复用方案；其分布近似匹配与贪心渐进匹配策略可迁移到任何需要低成本分布对齐的生成流程。
原文依据 we propose a Dual Matching Guided Diffusion (DMGD) framework, centered on efficient training-free guidance
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论数据集蒸馏通过将大规模数据集的信息蒸馏为规模显著更小的合成数据集，从而实现高效训练。近年来，基于扩散模型的范式不断涌现，为数据集蒸馏提供了新的视角。然而，这类方法通常需要额外的微调阶段，且有效的引导机制仍缺乏充分探索。为应对上述局限，我们重新审视基于扩散模型的数据集蒸馏，并提出双匹配引导扩散（DMGD）框架，其核心是高效的免训练引导。我们首先通过条件似然优化建立语义匹配，从而无需辅助分类器。进一步，我们提出一种动态引导机制，在保持语义对齐的同时提升合成数据的多样性。同时，我们引入基于最优传输（OT）的分布匹配方法，以进一步对齐目标分布结构。为保证效率，我们为基于扩散模型的框架设计了两种增强策略：分布近似匹配与贪心渐进匹配。这些策略以极小的计算开销实现有效的分布匹配引导。在 ImageNet-Woof、ImageNet-Nette 和 ImageNet-1K 上的实验结果表明，我们的免训练方法取得了显著提升，相较需要额外微调的最优方法，平均准确率分别提升 2.1%、5.4% 和 2.4%。
英文原文摘要 1531 字符
Dataset distillation enables efficient training by distilling the information of large-scale datasets into significantly smaller synthetic datasets. Diffusion based paradigms have emerged in recent years, offering novel perspectives for dataset distillation. However, they typically necessitate additional fine-tuning stages, and effective guidance mechanisms remain underexplored. To address these limitations, we rethink diffusion based dataset distillation and propose a Dual Matching Guided Diffusion (DMGD) framework, centered on efficient training-free guidance. We first establish Semantic Matching via conditional likelihood optimization, eliminating the need for auxiliary classifiers. Furthermore, we propose a dynamic guidance mechanism that enhances the diversity of synthetic data while maintaining semantic alignment. Simultaneously, we introduce an optimal transport (OT) based Distribution Matching approach to further align with the target distribution structure. To ensure efficiency, we develop two enhanced strategies for diffusion based framework: Distribution Approximate Matching and Greedy Progressive Matching. These strategies enable effective distribution matching guidance with minimal computational overhead. Experimental results on ImageNet-Woof, ImageNet-Nette, and ImageNet-1K demonstrate that our training-free approach achieves significant improvements, outperforming state-of-the-art (SOTA) methods requiring additional fine-tuning by average accuracy gains of 2.1%, 5.4%, and 2.4%, respectively.