AI Paper Atlas

Research map / Optimization / Implicit bias and generalization

Provable Guarantees for Nonlinear Feature Learning in Three-Layer Neural Networks

Eshaan Nichani, Alex Damian, Jason D. Lee

NeurIPS 2023 · Optimization › Implicit bias and generalization · approach: Neural tangent kernel analysis · 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 a three-layer network with layer-wise training instead of a two-layer network or kernel methodAlgorithm 1 · f*(x)=ReLU(x^T A x), d=32test loss vs sample sizeImproved
learn quadratic feature targets g*(x^T A x) with a three-layer network instead of kernel/NTK two-layer methodsAlgorithm 1 · functions of quadratic featuressample complexity: d^4 vs d^2pImproved
apply Algorithm 1 to single-index models with sigma2(z)=zAlgorithm 1 · single-index modelssample complexityNo clear effect
restrict link function g* to a degree p=O(1) polynomialAlgorithm 1 · g*(x^T A x)error floorImproved
use a two-layer network instead of Algorithm 1 on ReLU(x^T A x)two-layer network · depth separation targetwidth or weight norm requiredHurt
train layer-wise with Algorithm 1 on smoothed ReLU targetAlgorithm 1 · ReLU(x^T A x)-c0population L2 loss bound: d^-1/6 error floorImproved

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