AI Paper Atlas

Research map / Optimization / Implicit bias and generalization

Explicit loss asymptotics in the gradient descent training of neural networks

Maksim Velikanov, Dmitry Yarotsky

NeurIPS 2021 · Optimization › Implicit bias and generalization · approach: Neural tangent kernel analysis · 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
Increase network depth L from 1 to 3 hidden layersNTK-regime ReLU network with GP target, d=3 · Synthetic GP target, d=3loss exponent xi and eigenvalue exponent nuNo clear effect
Vary activation smoothness q in (z)_+^q (0.5, 0.75, 1.0, 1.5)Shallow NTK network · NTK eigenvalues, d=1 and d=3eigenvalue exponent nuMixed
Train in the mean-field regime instead of the NTK regimeShallow ReLU network · GP target, d=4eigenvalue exponent nuNo clear effect
Use only the singular part of the NTK instead of the full NTK (ablation)NTK eigenvalue spectrum · d = 1, 3, 6, 10; M = 100, 1000, 10000eigenvalues lambda_nNo clear effect
Increase input dimension dNTK eigenvalue convergence to asymptotic law · d = 1, 3, 6, 10index n at which spectrum matches asymptoticHurt
Switch target from GP-generated to indicator of a ballShallow ReLU NTK network · d=2, d=4loss exponent xi: 3/(d+1) -> 1/(d+1)Hurt
Apply theory to real data (MNIST)Relation xi = kappa/nu · MNISTfitted exponents: 0.26 ≈ 0.33/1.35Improved

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