AI Paper Atlas

Research map / Large language models / Language, safety and interpretability

DC-BENCH: Dataset Condensation Benchmark

Justin Cui, Ruochen Wang, Si Si, Cho-Jui Hsieh

NeurIPS 2022 · Large language models › Language, safety and interpretability · approach: Data and evaluation methodology · 13 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
enable best data augmentation at evaluation time (ablation)DC · CIFAR-10 IPC 10test accuracy (%): 44.43 -> 50.99Improved
enable best data augmentation at evaluation time (ablation)DSA · CIFAR-10 IPC 10test accuracy (%): 43.54 -> 52.96Improved
enable best data augmentation at evaluation time (ablation)DM · CIFAR-10 IPC 10test accuracy (%): 42.45 -> 47.64Improved
enable best data augmentation at evaluation time (ablation)TM · CIFAR-10 IPC 10test accuracy (%): 53.49 -> 63.66Improved
evaluate with augmentation instead of none (ablation)K-Center · CIFAR-100 IPC 50test accuracy (%): 33.61 -> 38.64Improved
K-Center selection vs DC synthesis under augmentationK-Center · CIFAR-100 IPC 50test accuracy (%): 30.56 -> 38.64 (CIFAR-100)Improved
increase IPC to 200 and abovecondensation methods vs random selection · CIFAR-10test accuracy (%)No clear effect
transfer condensed data from ConvNet to other architecturesTM · CIFAR-10 IPC 10test accuracy (%): 63.66 -> 31.55Hurt
use condensed data from DC for NAS ranking instead of full dataDC · NAS-Bench-201 CIFAR-10Spearman correlation: 0.7487 -> -0.19Hurt
use K-Center instead of random selection for NAS rankingK-Center · NAS-Bench-201 CIFAR-10Spearman correlation: -0.06 -> 0.11Improved
initialize synthetic images from K-Center instead of random selection (ablation)DM · CIFAR-10compute to same accuracy / final accuracy: about 30% budget; +1.3% finalImproved
scale ResNet152 training data from 300 to 1000 random images per classResNet152 · CIFAR-10test accuracy (%): 45.59 -> 70.29Improved

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