Dataset distillation (DD) condenses a large original dataset into a small distilled dataset with high training utility. Decoupled statistical matching methods substantially reduce distillation time and memory overhead while achieving strong performance. However, they typically supervise all distilled samples using runn...
Hong-Xu Ma, Guang Li, Shi-Jie Wang et al.· 0 citations
Medical image analysis often relies on large annotated datasets, whose collection, storage, and sharing are costly and constrained by privacy concerns. Dataset distillation offers a potential solution by constructing compact synthetic datasets that retain task-relevant information from the original data. However, gener...
Shuai-Ling Du, Guang Li, Ren Togo et al.· Bioengineering· 0 citations
ForUM is presented, a training-free test-time fusion of frozen MLLMs guided by two fixed geometric rules: agreement-based selection keeps the region supported by the most distinct models, and medoid localization returns an actual member box instead of a coordinate average, so one loose prediction cannot shift the answe...
Taiyo Sato, Takamasa Sanda, Keisuke Maeda et al.· 0 citations
This work asks whether the union of separately distilled sets reproduces training on the union of the real data composability and shows that it can fail even when every source is distilled exactly and the total budget admits an exact joint distillate.
DeCO uses attention rollout from a pretrained TransFG teacher to identify informative patches, applies spatial diversification to reduce redundant coverage, and organizes the resulting regions into class-wise evidence banks and consistently outperforms representative coreset and dataset-distillation baselines under dif...
Chuixuan Fan, Guang Li, Shi-Jie Wang et al.· 0 citations
Self-supervised representation-guided generative dataset distillation (SRG) is proposed, a framework that translates the SSL geometry into diffusion guidance and consistently outperforms the evaluated generative baselines across multiple datasets and IPC settings.
Mingzhuo Li, Guang Li, Linfeng Ye et al.· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
