AI Paper Atlas

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

Removing the Feature Correlation Effect of Multiplicative Noise

Zijun Zhang, Yining Zhang, Zongpeng Li

NeurIPS 2018 · Label-efficient and robust learning › Noisy labels · approach: Loss correction and noise robustness theory · 3 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
Replace standard multiplicative noise with NCMNCNN-16-10 · CIFAR-10test error (%): 3.76\pm 0.00 -> 3.41\pm 0.07Improved
Replace standard multiplicative noise with NCMNCNN-16-10 · CIFAR-100test error (%): 18.08\pm 0.03 -> 17.55\pm 0.06Improved
Replace standard multiplicative noise with NCMNWRN-22-7.5 · CIFAR-10test error (%): 3.59\pm 0.06 -> 3.02\pm 0.06Improved
Replace standard multiplicative noise with NCMNWRN-22-7.5 · CIFAR-100test error (%): 18.60\pm 0.03 -> 16.70\pm 0.13Improved
Use NCMN-0, which injects multiplicative noise directly into normalized signal features (ablation)NCMN · CIFAR-10, CNN-16-10test error (%): 3.41\pm 0.07 -> 3.51\pm 0.07Improved
Use NCMN-0 instead of NCMN-1 (ablation)NCMN · CIFAR-100, CNN-16-10test error (%): 17.55\pm 0.06 -> 17.37\pm 0.05Improved
Apply NCMN noise once every two layers (NCMN-2) instead of every layer (ablation)NCMN · CIFAR-10, CNN-16-10test error (%): 3.41\pm 0.07 -> 3.44\pm 0.03No clear effect
Apply NCMN noise once every two layers (NCMN-2) instead of every layer (ablation)NCMN · CIFAR-100, CNN-16-10test error (%): 17.55\pm 0.06 -> 18.16\pm 0.04Hurt
Use NCMN-2 instead of NCMN-1 (ablation)NCMN · CIFAR-10, WRN-22-7.5test error (%): 3.02\pm 0.06 -> 3.00\pm 0.05No clear effect
Use NCMN-2 instead of NCMN-1 (ablation)NCMN · CIFAR-100, WRN-22-7.5test error (%): 17.09\pm 0.10 -> 16.70\pm 0.13Improved
Replace dropout with NCMNLSTM language models · LSTM experimentsfeature correlationNo 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