AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

Dataset Distillation using Neural Feature Regression

Yongchao Zhou, Ehsan Nezhadarya, Jimmy Ba

NeurIPS 2022 · Architectures and efficiency › RNNs and normalization · approach: Dataset distillation and data-efficient training · 33 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Use FRePo rather than MTT for one-image-per-class distillationDataset distillation · CIFAR100, 1 Img/Clstest accuracy: 24.3 ± 0.3 -> 28.7 ± 0.1Improved
Use FRePo rather than MTT for one-image-per-class distillationDataset distillation · Tiny ImageNet, 1 Img/Clstest accuracy: 8.8 ± 0.3 -> 15.4 ± 0.3Improved
Use FRePo rather than MTT for one-image-per-class distillationDataset distillation · CUB-200, 1 Img/Clstest accuracy: 2.2 ± 0.1 -> 12.4 ± 0.2Improved
Use FRePo rather than MTT for 50 images per classDataset distillation · CIFAR100, 50 Img/Clstest accuracy: 47.7 ± 0.2 -> 44.3 ± 0.2Hurt
Use FRePo rather than MTT for 10 images per classDataset distillation · CIFAR100, 10 Img/Clstest accuracy: 40.1 ± 0.4 -> 42.5 ± 0.2Improved
Use FRePo instead of a random real subsetImageNet distillation · ImageNet (64x64), 1 Img/ClsTop1 accuracy: 1.1 ± 0.1 -> 7.5 ± 0.3Improved
Use a model pool rather than only one online model (ablation)FRePo · CIFAR100, 1 Img/Clstest accuracyImproved
Use an intermediate number of online updates rather than too few or too many (ablation)FRePo · CIFAR100, 1 Img/Clstest accuracyMixed
Increase meta-gradient batch size (ablation)FRePo · CIFAR100, 1 Img/Clstest accuracyImproved
Use TBPTT instead of the FRePo kernel-regression meta-gradient (ablation)Dataset distillation · CIFAR100, 1 Img/Clstest accuracyHurt
Compute the meta-gradient before rather than after the online model update (ablation)FRePo · CIFAR100, 1 Img/Clstest accuracyNo clear effect

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy