AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

Globally Gated Deep Linear Networks

Qianyi Li, Haim Sompolinsky

NeurIPS 2022 · Architectures and efficiency › RNNs and normalization · approach: Theory of depth and width · 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
Use finite-width kernel shape renormalization instead of the GP kernelGGDLN · MNISTgeneralization errorImproved
Use pretrained unsupervised soft k-means gatings instead of random gatingsGGDLN · MNISTerror rateImproved
Increase the number of gatingsGGDLN · MNISTerror rateImproved
Increase network widthGGDLN with localized receptive fields · ReLU teacher taskgeneralization errorImproved
Increase network widthGGDLN with localized receptive fields · ReLU teacher task with preferred inputsgeneralization errorHurt
Increase regularization amplitude sigma in the regime where bias falls and variance risesGGDLN · MNISTerror rateMixed
Increase regularization amplitude sigma in the regime where bias and variance both riseGGDLN · MNISTerror rateHurt
Use finite-width kernel renormalizationDeep GGDLN · Deep GGDLN taskbiasImproved
Increase network widthGGDLN with bottom-up gatings · Permuted MNISTerror rateImproved
Increase gating thresholdGGDLN with bottom-up gatings · Permuted MNISTerror rateMixed
Increase network widthGGDLN with combined top-down and bottom-up gatings · Permuted and un-permuted MNIST tasksgeneralization errorHurt
Increase network depthGGDLN with renormalized kernel · Deep GGDLN taskkernel flatteningImproved

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