无需对齐外部编码器，直接让扩散模型学出可线性分离的表征。
No Alignment Needed for Generation: Learning Linearly Separable Representations in Diffusion Models
- 通过引入线性可分性正则化，替代传统特征对齐策略。
- 在256×256 ImageNet上实现FID 1.46，生成质量显著提升。
- 适用于追求高效训练与高质量生成的扩散模型研究者。
大规模扩散模型的高效训练近年来聚焦于提升其判别性表征能力。主流方法依赖外部预训练编码器进行表征对齐，虽能提升线性探测性能，但需昂贵的预训练编码器。本文提出一种新正则化策略——线性可分性（LSEP），无需辅助编码器和特征对齐，将线性探测直接融入网络学习过程。实验表明，在基于流的Transformer架构如SiTs上，该方法显著提升训练效率与生成质量，256×256 ImageNet上的FID达1.46。
原文摘要 · Abstract (English)
Efficient training strategies for large-scale diffusion models have recently emphasized the importance of improving discriminative feature representations in these models. A central line of work in this direction is representation alignment with features obtained from powerful external encoders, which improves the representation quality as assessed through linear probing. Alignment-based approaches show promise but depend on large pretrained encoders, which are computationally expensive to obtain. In this work, we propose an alternative regularization for training, based on promoting the Linear SEParability (LSEP) of intermediate layer representations. LSEP eliminates the need for an auxiliary encoder and representation alignment, while incorporating linear probing directly into the network's learning dynamics rather than treating it as a simple post-hoc evaluation tool. Our results demonstrate substantial improvements in both training efficiency and generation quality on flow-based transformer architectures such as SiTs, achieving an FID of 1.46 on $256 \times 256$ ImageNet dataset.
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。