AI Paper Atlas

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

A Lazy Approach to Long-Horizon Gradient-Based Meta-Learning

Muhammad Abdullah Jamal, Liqiang Wang, Boqing Gong

ICCV 2021 · Label-efficient and robust learning › Knowledge distillation · approach: Teacher properties and robustness · 9 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
add lazy teacher leap with long inner horizonMAML · MiniImageNet 5-way 5-shotaccuracy (%): 63.11 -> 64.13Mixed
add lazy teacher leap with long inner horizonMAML · MiniImageNet 5-way 1-shotaccuracy (%): 48.70 -> 48.26No clear effect
add lazy teacher leap with long inner horizonReptile · MiniImageNet 5-way 1-shotaccuracy (%): 49.97 -> 51.50Improved
use last 3 checkpoints instead of 1 in the convex hull (ablation)Lazy Reptile · MiniImageNet 5-way 5-shotaccuracy (%): 67.22 -> 68.77Improved
use last 3 checkpoints instead of 1 in the convex hull (ablation)Lazy MAML · MiniImageNet 5-way 1-shotaccuracy (%): 48.26 -> 48.17No clear effect
add lazy teacher to MAML with more shotsMAML · MiniImageNet 5-way 50-shotaccuracy (%): 67.13 -> 78.54Improved
add lazy teacher to MAML with more shotsMAML · MiniImageNet 5-way 20-shotaccuracy (%): 66.18 -> 74.8Improved
add lazy teacher to MAML on non-i.i.d. tasks (no class shuffling)MAML · Omniglot 20-way 1-shot (non-i.i.d.)accuracy (%): 81.51 -> 86.35Improved
use Reptile-style update instead of MAML-style in lazy method (ablation)Lazy MAML · MiniImageNet 5-way 50-shotaccuracy (%): 78.54 -> 76.17Hurt
add lazy teacher to two-component weightingTwo-component weighting · CIFAR-LT-100 (imbalance 200)top-1 error (%): 60.69 -> 58.67Improved
add lazy teacher to two-component weightingTwo-component weighting · CIFAR-LT-100 (imbalance 20)top-1 error (%): 44.38 -> 43.68Improved
add lazy teacher to meta-attackMeta-attack · CIFAR10 untargeted attackavg queries: 1,583 -> 1,061Improved

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