AI Paper Atlas

Research map / Label-efficient and robust learning / Knowledge distillation

IDEAL: Query-Efficient Data-Free Learning from Black-Box Models

Jie Zhang, Chen Chen, Lingjuan Lyu

ICLR 2023 · Label-efficient and robust learning › Knowledge distillation · approach: Distillation with causal and adaptive guidance · 5 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 the student model, rather than the teacher, to train the generator without querying the teacherIDEAL · CIFAR10top-1 test accuracy (%): 68.82 vs 37.91Improved
Use the student model, rather than the teacher, to train the generator without querying the teacherIDEAL · CIFAR100top-1 test accuracy (%): 36.96 vs 20.01Improved
Add the information loss L_infor to generator training (ablation)IDEAL · CIFAR10top-1 test accuracy (%): 62.68 -> 68.82Improved
Add the information loss L_infor to generator training (ablation)IDEAL · SVHNtop-1 test accuracy (%): 83.21 -> 86.91Improved
Add the cross-entropy loss L_ce to generator training (ablation)IDEAL · CIFAR10top-1 test accuracy (%): 10.68 -> 68.82Improved
Add the cross-entropy loss L_ce to generator training (ablation)IDEAL · SVHNtop-1 test accuracy (%): 12.48 -> 86.91Improved
Reinitialize the generator at each epoch (ablation)IDEAL · CIFAR10top-1 test accuracy (%): 55.32 -> 68.82Improved
Reinitialize the generator at each epoch (ablation)IDEAL · SVHNtop-1 test accuracy (%): 80.53 -> 86.91Improved
Set generator-training iterations E_G to 3 instead of the reported best value 5 (ablation)IDEAL · MNISTtop-1 test accuracy (%): 65.36 vs 96.51Hurt
Set generator-training iterations E_G to 50 instead of the reported best value 10 (ablation)IDEAL · CIFAR10top-1 test accuracy (%): 68.82 -> 58.69Hurt
Use larger generator-training iteration counts (ablation)IDEAL · CIFAR10top-1 test accuracy (%): 30.25, 57.56, 68.82, 62.45, 58.69Mixed
Use generator re-initialization and balanced per-class sample generation rather than class-imbalance reweighting (ablation)IDEAL · CIFAR10top-1 test accuracy (%): 62.14, 63.31 -> 68.82Improved

Later papers that compared with this method

2 later accepted papers report IDEAL as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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