A generative dataset distillation method that replaces MSE logits matching with standardized KL-divergence distribution matching, reporting improved accuracy on three benchmarks.
Applica- tions for deep learning in ecology,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Generative Dataset Distillation Based on Self-knowledge Distillation
A generative dataset distillation method that replaces MSE logits matching with standardized KL-divergence distribution matching, reporting improved accuracy on three benchmarks.