AI Paper Atlas

Research map / Learning theory and algorithms / Prediction and decision losses

Learning to Learn with Smooth Regularization

Yuanhao Xiong, Cho-Jui Hsieh

ECCV 2022 · Learning theory and algorithms › Prediction and decision losses · approach: Losses for decision-making · 0 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 smoothness-inducing regularizerSimpleOptimizer · CIFAR-10 3-layer CNNtest accuracy (%): 69.02 -> 72.50Improved
use PGD smoothness regularizer instead of l2 regularization (ablation)SimpleOptimizer · CIFAR-10 3-layer CNNtest accuracy (%): 69.69 -> 72.50Improved
use PGD smoothness regularizer instead of spectral normalization (ablation)SimpleOptimizer · CIFAR-10 3-layer CNNtest accuracy (%): 69.35 -> 72.50Improved
add orthogonal regularization (ablation)SimpleOptimizer · CIFAR-10 3-layer CNNtest accuracy (%): 69.02 -> 10.00Hurt
add l2 regularization (ablation)SimpleOptimizer · CIFAR-10 3-layer CNNtest accuracy (%): 69.02 -> 69.69No clear effect
add smoothness regularizerMeta-LSTM · miniImageNet 5-way 1-shotaccuracy (%): 38.20 -> 40.42Improved
add smoothness regularizerMeta-LSTM · miniImageNet 5-way 5-shotaccuracy (%): 56.56 -> 58.90Improved
add smoothness regularizerMeta-LSTM · tieredImageNet 5-way 1-shotaccuracy (%): 36.43 -> 36.74No clear effect
add smoothness regularizerMeta-LSTM · tieredImageNet 5-way 5-shotaccuracy (%): 53.45 -> 55.14Improved
add smoothness regularizerSIB WRN-28-10 · miniImageNet 5-way 5-shotaccuracy (%): 78.9 -> 80.8Improved
add smoothness regularizerSIB WRN-28-10 · miniImageNet 5-way 1-shotaccuracy (%): 69.6 -> 70.0No clear effect
add smoothness regularizerSIB WRN-28-10 · CIFAR-FS 5-way 5-shotaccuracy (%): 85.3 -> 86.1Improved

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