EVLF：面向生成式数据集蒸馏的早期视觉-语言融合
EVLF: Early Vision-Language Fusion for Generative Dataset Distillation
提出在编码器与生成主干之间插入轻量交叉注意力实现早期视觉-语言融合，可即插即用地改进任何基于扩散的数据集蒸馏流程，为视觉研究者提供了一种平衡文本语义与视觉潜变量的生成控制手段。
原文依据 we introduce an Early Vision-Language Fusion (EVLF) method that aligns textual and visual embeddings at the transition between the encoder and the generative backbone
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（DD）旨在合成紧凑的训练集，使模型能够以显著更少的样本达到高准确率。近期基于扩散的DD方法通常通过后期交叉注意力引入语义引导，其中文本提示词往往主导生成过程。尽管该策略强制了标签相关性，但它削弱了视觉潜变量的贡献，导致样本被过度校正，反映的是提示词模式而非内在视觉特征。为解决此问题，我们提出早期视觉-语言融合（EVLF）方法，在编码器与生成主干之间的过渡处对齐文本与视觉嵌入。通过在此过渡处加入轻量级交叉注意力模块，早期表示在整个去噪过程中同时编码局部纹理和全局语义方向。重要的是，EVLF即插即用，可轻松集成到任何带有编码器的基于扩散的数据集蒸馏流程中。它适用于不同的去噪器架构和采样调度，无需任何任务特定的修改。大量实验表明，EVLF生成语义忠实且视觉连贯的合成数据，在不同设置下均能持续提升下游分类准确率。源代码见 https://github.com/wenqi-cai297/earlyfusion-for-dd 。
英文原文摘要 1455 字符
Dataset distillation (DD) aims to synthesize compact training sets that enable models to achieve high accuracy with significantly fewer samples. Recent diffusion-based DD methods commonly introduce semantic guidance through late-stage cross-attention, where textual prompts tend to dominate the generative process. Although this strategy enforces label relevance, it diminishes the contribution of visual latents, resulting in over-corrected samples that mirror prompt patterns rather than reflecting intrinsic visual features. To solve this problem, we introduce an Early Vision-Language Fusion (EVLF) method that aligns textual and visual embeddings at the transition between the encoder and the generative backbone. By incorporating a lightweight cross-attention module at this transition, the early representations simultaneously encode local textures and global semantic directions across the denoising process. Importantly, EVLF is plug-and-play and can be easily integrated into any diffusion-based dataset distillation pipeline with an encoder. It works across different denoiser architectures and sampling schedules without any task-specific modifications. Extensive experiments demonstrate that EVLF generates semantically faithful and visually coherent synthetic data, yielding consistent improvements in downstream classification accuracy across varied settings. Source code is available at https://github.com/wenqi-cai297/earlyfusion-for-dd .