Method · Synthetic-set parameterization
Translative pre-training
Few-Shot Dataset Distillation via Translative Pre-Training
Songhua Liu, Xinchao Wang
ICCV 2023 · first public 2023-01
In one paragraph
Learns a 'distillation space' via a translator network, pretrained on large datasets with image-to-image translation, that maps synthetic images optimized cheaply in an arbitrary fixed network's space into the space of the target few-shot distillation network, needing only a few adaptation steps per new target dataset; reports about 15x faster distillation with performance comparable to iterative bi-level baselines, and satisfactory generalization across datasets, budgets and class counts.
Where it sits
- Synthetic-set parameterization (Orthogonal design choices)
- Optimization and training recipes (Orthogonal design choices)
- Setting: Image classification
Design choices
| What is stored | other |
BibTeX (generated; prefer the venue's official entry)
@article{liu2023shot,
title = {Few-Shot Dataset Distillation via Translative Pre-Training},
author = {Songhua Liu and Xinchao Wang},
journal = {ICCV 2023},
year = {2023}
}Nearby in Synthetic-set parameterization
Post Training Quantization for Efficient Dataset Condensation
Linh-Tam Tran, Sung-Ho Bae · AAAI 2026notablepaper ↗
3DDP — Parameterization-Based Dataset Distillation of 3D Point Clouds through Learnable Shape Morphing
Dongwook Kim, Jae-Young Yim · ICLR 2026notableOther datapaper ↗code ↗
Rate-utility DD — Dataset Distillation as Data Compression: A Rate-Utility Perspective
Youneng Bao, Yiping Liu, Zhuo Chen et al. · ICCV 2025notablepaper ↗code ↗