Dataset Distillation via Drifting：不用分类器，直接在潜空间逼近真实分布
Dataset Distillation via Drifting
搞数据集蒸馏的人可以看看这篇：不靠分类器优化，直接在潜空间里把合成样本往真实分布上漂移，分布覆盖比老方法好。还顺带送一个模型可视化解释的用法，图像之外的模态也能跑。
ArXiv 论文（编号 2611.16642）提出 Dataset Distillation via Drifting，一种新的数据集蒸馏方法。传统蒸馏大多围绕一个分类器来优化合成样本，而该方法改以特征提取器的潜空间为载体，让合成样本同时向真实数据靠近、彼此远离，全程不使用分类目标。
在分布覆盖上，drifted 数据对真实分布的拟合显著优于既有方法，同时作为分类训练样本仍保持强竞争力。作者还展示了该方法可作模型解释的可视化工具，并从图像扩展到其他模态，附带提出基于分布的互补评估指标。
Dataset Distillation via Drifting
The task of dataset distillation concerns the learning of a small set of synthetic samples that can serve as an effective replacement for a large real dataset when used for downstream tasks. While many solutions have been proposed since the problem's inception, most of them primarily seek to optimize the distilled data with respect to a classifier. In this work, we propose a new distillation method that learns a set of samples such that they directly approximate the per-class data distribution. Specifically, we "drift" our distilled data in the latent space of a feature extractor such that they move both toward the real data and away from each other without using a classification objective. We find that our drifted data cover the real distribution significantly better than existing methods while still serving as powerful training samples for classification. Furthermore, we showcase our method's use as a visualization tool for better interpreting models and extend it beyond images to other modalities. Our work advances the frontier of dataset distillation, offering a novel synthesis method and proposing distribution-based metrics as a complementary evaluation. More broadly, our findings help address the growing need to understand both data and models by uncovering more about how they interact.
本条情报归属 AITOP「论文」核心追踪栏目。涉及核心实体与技术议题包括：数据集蒸馏、Dataset Distillation、合成数据、分布匹配，读者可通过下方标签进入对应专题页查阅历史脉络与关联对比。
信源核验说明：本文原始线索采编自「arXiv cs.AI」（@George Cazenavette, Angelina Quan, Giannis Daras, Antonio Torralba, Vincent Sitzmann），经 AITOP 多维评分管线结构化提炼。