AI Paper Atlas

Research map / Label-efficient and robust learning / Noisy labels

Retaining Beneficial Information from Detrimental Data for Neural Network Repair

Long-Kai Huang, Peilin Zhao, Junzhou Huang, Sinno Pan

NeurIPS 2023 · Label-efficient and robust learning › Noisy labels · approach: Sample selection and co-training · 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
align identified data with energy-based alignment instead of deleting it (BIR vs EWC-Deletion), imprecise cause setmodel treatment, LeNet · MNIST-Label (75% cause + 25% clean)test accuracy (All, %): 87.74 -> 95.70Improved
use imprecise rather than precise cause set for BIRBIR, LeNet · MNIST-Labeltest accuracy (All, %): 95.64 -> 95.70No clear effect
use imprecise rather than precise cause set for EWC-DeletionEWC-Deletion, LeNet · MNIST-Labeltest accuracy (All, %): 95.10 -> 87.74Hurt
RDIA relabeling of identified dataRDIA, LeNet · MNIST-Labeltest accuracy (All, %): 82.46 -> 72.70Hurt
apply BIR repair to the base modelLeNet · MNIST-Labeltest accuracy (All, %): 82.46 -> 95.60Improved
apply BIR repair to the base modelResNet50 · CIFAR10-Advtest accuracy (Noisy classes, %): 79.30 -> 85.85Improved
apply SGD Influence Repair to the base modelResNet50 · CIFAR10-Advtest accuracy (All, %): 87.60 -> 84.08Hurt
apply BIR repair to the base modelLeNet · MNIST-Inputtest accuracy (All, %): 96.46 -> 96.16Hurt
apply BIR repair to the base modelResNet20 · CIFAR-10N Worsttest accuracy (%): 77.76 -> 85.50Improved
apply SGD Influence Repair to the base modelResNet20 · CIFAR-10N Worsttest accuracy (%): 77.76 -> 71.50Hurt
update only Layer 1 instead of the selected layers (ablation)BIR, ResNet20 · CIFAR10-Labeltreatment accuracy (%): 83.98 -> 74.82Hurt
update all layers instead of the selected layer subset (ablation)BIR, LeNet · MNIST-Labeltreatment accuracy (%): 95.60 -> 95.50No 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